A high-precision gas detection adaptive light intensity noise suppression system and method

Abstract

This invention relates to the field of gas concentration monitoring technology, providing a high-precision gas detection adaptive light intensity noise suppression system and method. The system includes a detection module, a gas input module, and an adjustment module. The system has a compact structure and not only effectively suppresses multi-scale light intensity noise introduced by the laser source, but also significantly improves detection precision while ensuring rapid system response. It employs a coaxial optical path design to sense light intensity noise and combines it with a dual-timescale noise feature extraction method to achieve adaptive light intensity noise suppression. Maintaining high precision while possessing rapid response capabilities, it provides a highly reliable detection method for scenarios such as oil and gas exploration and safe production.

Description

Technical Field

[0001] This invention belongs to the field of gas concentration monitoring technology, and particularly relates to a high-precision gas detection adaptive light intensity noise suppression system and method. Background Technology

[0002] Gas concentration monitoring is crucial for oil and gas exploration, emissions regulation, and industrial safety. Therefore, a high-precision technology is urgently needed to provide excellent detection performance. Tunable diode laser absorption spectroscopy (TDLAS) is favored for its high sensitivity and excellent specificity, making it well-suited to meet the requirements of these applications. Particularly in the mid-infrared region, methane exhibits a strong fundamental absorption band, with a significantly enhanced signal compared to the near-infrared region. Theoretically, this further strengthens its advantages, suggesting even higher performance. However, in practical applications, the inherent instability of the laser source itself remains a significant challenge, limiting the accuracy of long-term gas measurements.

[0003] Quantum cascade lasers (QCLs) have become a widely used light source in mid-infrared TDLAS systems. However, fundamental device physics issues such as temperature sensitivity, drive current noise, and material inhomogeneities can lead to spectral drift, intensity fluctuations, and beam quality degradation, resulting in severe intensity noise. This noise originates from the inherent characteristics of the laser itself, exhibiting a random and non-stationary process that cannot be fully described or predicted by deterministic models. Crucially, in TDLAS systems, this noise is multiplicatively coupled to the absorption signal. It is not simply superimposed on the signal but directly disguised as the absorption signal, completely masking the true concentration information. This inherent indistinguishability and unpredictability cause significant fluctuations and severe drift in TDLAS systems during long-term operation, fundamentally limiting their reliability in high-precision applications.

[0004] Current techniques for suppressing optical intensity noise in TDLAS systems still have significant limitations. Regarding active laser frequency control, while feedback calibration methods can alleviate wavelength drift, they are bulky and costly. Furthermore, the controller struggles to balance high gain and low noise across a wide frequency range, failing to simultaneously suppress rapid fluctuations and slow drift. Current adjustment can also induce thermal instability and optical power fluctuations, introducing additional measurement errors. While common-mode suppression methods based on parallel reference channels are widely used, their reference signals are affected not only by laser characteristics but also by mechanical disturbances from beam splitters, thermal deformation of optical components, and circuit phase drift, making them ineffective at decoupling non-common-mode noise and potentially introducing new signal distortions. In addition, calibration techniques such as harmonic ratio methods are limited by absorbance range, limiting their applicability. At the signal processing level, traditional filtering algorithms (such as Kalman filtering and wavelet denoising) rely on data stationarity or prior noise statistical assumptions, making them ill-suited for real-world non-stationary noise processes. While hybrid algorithms and emerging artificial intelligence methods have improved the ability to handle complex noise, they still have problems such as model dependence on large amounts of data, high computational complexity, and weak generalization ability, making it impossible to balance long-term stability and real-time requirements.

[0005] In summary, existing methods have shortcomings in dealing with multi-scale, non-stationary, and signal-coupled light intensity noise, which restricts further improvement of the measurement accuracy of TDLAS systems. Summary of the Invention

[0006] The purpose of this invention is to provide a high-precision gas detection adaptive light intensity noise suppression system and method, which aims to solve the problems mentioned in the background art.

[0007] The present invention is implemented as follows: a high-precision gas detection adaptive light intensity noise suppression system includes a detection module, a gas input module, and an adjustment module;

[0008] The detection module includes a measurement gas chamber and a quantum cascade laser. The light source emitted by the quantum cascade laser passes through a reference gas chamber and a plano-convex lens in sequence before entering the measurement gas chamber, and finally exits from the measurement gas chamber and enters the detector.

[0009] The gas input module includes a first proportional valve and a vacuum pump. The first proportional valve is connected to a mass flow meter, and the mass flow meter is connected to the gas inlet of the measuring gas chamber. The vacuum pump is connected to a second proportional valve, the other end of which is connected to the gas outlet of the measuring gas chamber. A pressure sensor is installed on the pipeline between the second proportional valve and the measuring gas chamber.

[0010] The adjustment module includes a pressure control unit, a host computer, a signal generator, and a temperature control unit. The pressure control unit is connected to a first proportional valve, a mass flow meter, a pressure sensor, a second proportional valve, and a vacuum pump. The signal generator is connected to a quantum cascade laser and is also connected to a lock-in amplifier. The detector is connected to the lock-in amplifier, and the lock-in amplifier is also connected to a data acquisition card. The host computer is connected to the pressure control unit, the data acquisition card, and the temperature control unit.

[0011] In a further technical solution, the reference gas chamber is filled with a reference gas whose absorption spectrum does not cross-interfere with the measurement gas; the reference gas chamber adopts a wedge-shaped structure.

[0012] In a further technical solution, the pressure control unit is used to control the system pressure in the measuring chamber at 30 Torr, with fluctuations within ±0.02 Torr; the temperature control unit is used to control the system temperature at 22℃, with fluctuations within ±0.01℃.

[0013] Another objective of this invention is to provide a high-precision gas detection adaptive light intensity noise suppression method, based on the above-described system, comprising the following steps:

[0014] Step 1: Drive the quantum cascade laser to emit laser light through the signal generator. The laser light passes through the reference gas cell and the measurement gas cell in sequence and is received by the detector. After being demodulated by the lock-in amplifier, the second harmonic signal amplitude of the reference gas and the measurement gas is collected by the data acquisition card and transmitted to the host computer.

[0015] Step 2: Define the system state vector and observation vector, and establish the state transition equation and measurement equation;

[0016] Step 3: Construct a dual-timescale noise feature extraction strategy to decompose the perceived light intensity noise signal into drift features and fluctuation features;

[0017] Step 4: Construct an adaptive estimation model to predict and update the system state and error covariance, and use a monitor-reset strategy to enhance the tracking response capability. Finally, output the measured gas concentration value with suppressed light intensity noise.

[0018] A further technical solution, in step 1, according to the Lambert-Beer law, when using TDLAS technology based on coaxial sensing for gas detection, the relationship between incident light intensity and transmitted light intensity is expressed as:

[0019] (1);

[0020] in, It is the initial light intensity. It measures light intensity. It is a natural exponential function, representing the exponential decay of light intensity. It refers to the intensity of the reference gas absorption line. It is a reference gas absorption line shape. It is the reference gas pressure. It is a reference gas concentration. It is the reference gas absorption optical path. It measures the intensity of gas absorption lines. It measures the shape of the gas absorption line. It measures gas pressure. It measures gas concentration. It measures the optical path length of gas absorption;

[0021] In practice, due to the presence of laser noise, the laser intensity model is as follows:

[0022] (2);

[0023] in, The initial light intensity under ideal conditions; This represents the intensity noise of the laser beam, which is a non-stationary random process that follows a certain order. distributed; It is the distribution function; This represents a time-varying mean function, indicating the drift of the laser power; This represents a time-varying variance function, indicating the rapid fluctuations in laser power;

[0024] Therefore, within one scan cycle, the light intensities received by the detector at the reference time and the measurement time are respectively:

[0025] (3);

[0026] in, To reference the laser scanning time corresponding to the peak absorption of the gas, To measure the laser scanning time corresponding to the gas absorption peak, To reference the initial laser intensity corresponding to the peak absorption of the gas, To measure the initial laser intensity corresponding to the gas absorption peak, a lock-in amplifier was used to extract the second harmonic signal of each absorption peak:

[0027] (4);

[0028] in, The amplitude of the second harmonic signal of the reference gas; To measure the amplitude of the second harmonic signal of the gas; The system gain for the reference gas; To measure the system gain of the gas.

[0029] In a further technical solution, step 2 includes the following specific steps:

[0030] The system state vector and observation vector are defined as follows:

[0031] (5);

[0032] in, This is the system state vector; This represents the system observation vector; express The gas concentration is measured continuously and expressed in units of voltage. This is the light intensity drift coefficient. ; for laser light intensity at any given time; Nominal light intensity; The measured value is the second harmonic amplitude of the gas. The measured reference gas second harmonic amplitude;

[0033] Establish the following state transition equations and measurement equations:

[0034] (6);

[0035] in, Represents the state transition matrix; Represents the observation matrix; This is process noise; For measuring noise.

[0036] A further technical solution involves, in step 3, using the reference air cell signal to calculate the residual of the reference signal in order to extract noise fluctuations and adaptively adjust the noise matrix; the estimated measurement noise variance is expressed as:

[0037] (7);

[0038] (8);

[0039] in, This represents the estimated variance of the measurement noise; It is the average amplitude of the second harmonic of the reference gas within the window; Indicates the scaling factor; This represents the size of the sliding window used to estimate the variance; This represents the forgetting factor.

[0040] To extract the drift trend, a weighted moving average method was used to eliminate the influence of short-term laser intensity fluctuations; the trend characteristics of the reference signal are represented as follows:

[0041] (9);

[0042] in, Indicates the laser drift trend; This represents the sliding window size for drift trend extraction; the observation matrix is... Measurement noise covariance matrix The update is as follows:

[0043] .

[0044] A further technical solution involves predicting the system state and error covariance in step 4:

[0045] (10);

[0046] (11);

[0047] in, The predicted value of the state vector; This represents the predicted value of the error covariance matrix at the current time. This represents the predicted value of the error covariance matrix at the previous time step; Represents the process noise covariance matrix; state transition matrix ,in This is the gas concentration diffusion correction coefficient; a monitoring-reset strategy is adopted to enhance the tracking response capability of the algorithm. This strategy comprehensively considers current and past information data to determine whether to reset the estimated state; the decision criteria are set as follows:

[0048] (12);

[0049] (13);

[0050] in, Represents the innovation vector. Indicates the monitoring threshold. Indicates the length of the historical information sequence;

[0051] Calculate the gain matrix:

[0052] (14);

[0053] in, Represents the observation matrix. This represents the predicted value of the error covariance matrix;

[0054] Status Update:

[0055] (15);

[0056] Error covariance update:

[0057] (16);

[0058] in, This represents the predicted value of the error covariance matrix at the next time step. To robustly adjust the coefficients, It is an identity matrix.

[0059] This invention provides a high-precision gas detection adaptive light intensity noise suppression system and method. While maintaining a compact system structure, it effectively suppresses multi-scale light intensity noise introduced by the laser source and significantly improves detection precision while ensuring rapid system response. The system employs a coaxial optical path design to sense light intensity noise and combines it with a dual-timescale noise feature extraction method to achieve adaptive light intensity noise suppression. Long-term measurement experimental data show that the adaptive light intensity noise suppression method improves the long-term stability of the detection system by 7.4 times. Under conditions of a short optical path of 5 meters and an integration time of 0.5 seconds, the system can achieve a detection precision of 0.017 ppmv, an improvement of more than 100 times compared to the original signal, effectively supporting continuous and reliable monitoring. Maintaining high precision while possessing rapid response capabilities, it provides a highly reliable detection method for scenarios such as oil and gas exploration and safe production. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of a high-precision gas detection adaptive light intensity noise suppression system provided in an embodiment of the present invention;

[0061] Figure 2 A flowchart of a high-precision gas detection adaptive light intensity noise suppression method provided in an embodiment of the present invention;

[0062] Figure 3 This is a long-term measurement result;

[0063] Figure 4 This is a graph of Allen's bias data.

[0064] In the attached diagram: 1. First proportional valve; 2. Mass flow meter; 3. Pressure control unit; 4. Pressure sensor; 5. Second proportional valve; 6. Vacuum pump; 7. Host computer; 8. Detector; 9. Measuring gas chamber; 10. Reference gas chamber; 11. Plano-convex lens; 12. Quantum cascade laser; 13. Signal generator; 14. Lock-in amplifier; 15. Data acquisition card; 16. Temperature control unit. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0066] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0067] like Figure 1 As shown, a high-precision gas detection adaptive light intensity noise suppression system according to an embodiment of the present invention includes a detection module, a gas input module, and an adjustment module.

[0068] The detection module includes a measurement gas chamber 9 and a quantum cascade laser 12. The light source emitted by the quantum cascade laser 12 passes through the reference gas chamber 10 and the plano-convex lens 11 in sequence before entering the measurement gas chamber 9, and finally exits from the measurement gas chamber 9 and enters the detector 8.

[0069] The gas input module includes a first proportional valve 1 and a vacuum pump 6. The first proportional valve 1 is connected to a mass flow meter 2, and the mass flow meter 2 is connected to the gas inlet end of the measuring gas chamber 9. The vacuum pump 6 is connected to a second proportional valve 5, and the other end of the second proportional valve 5 is connected to the gas outlet of the measuring gas chamber 9. A pressure sensor 4 is installed on the pipeline between the second proportional valve 5 and the measuring gas chamber 9.

[0070] The adjustment module includes a pressure control unit 3, a host computer 7, a signal generator 13, and a temperature control unit 16. The pressure control unit 3 is connected to a first proportional valve 1, a mass flow meter 2, a pressure sensor 4, a second proportional valve 5, and a vacuum pump 6. The signal generator 13 is connected to a quantum cascade laser 12 and is also connected to a lock-in amplifier 14. The detector 8 is connected to the lock-in amplifier 14, and the lock-in amplifier 14 is also connected to a data acquisition card 15. The host computer 7 is connected to the pressure control unit 3, the data acquisition card 15, and the temperature control unit 16.

[0071] In this embodiment of the invention, the system is used to achieve high-precision measurement of methane gas, employing a quantum cascade laser 12 with a center wavelength of 7.71 µm as the light source. By adjusting the injection current of the quantum cascade laser 12 within the range of 279 mA to 295 mA, the wavenumber of the output laser can be continuously tuned to 1296.070 cm⁻¹. -1 Up to 1296.225 cm -1 Within this range, 1296.1182 cm is selected. -1 The methane absorption line at 1296.1595 cm⁻¹ was selected as the measurement object.-1 The acetylene (C2H2) absorption line is used as a stable reference line. The absorption coefficients of the two absorption lines differ by approximately 1500 times, and their mutual interference is negligible. The optical path of the reference gas cell 10, filled with C2H2, is 3 cm, and the acetylene pressure is 5 Torr. The reference gas cell 10 adopts a wedge-shaped structure design to avoid interference noise from the parallel plates and to prevent some feedback light from entering the laser.

[0072] In the optical path, the laser emitted from the quantum cascade laser 12 enters the reference gas chamber 10, is focused by the plano-convex lens 11 into the measuring gas chamber 9, and the emitted light from the measuring gas chamber 9 is received by the detector 8. After demodulation by the lock-in amplifier 14, it is acquired by the data acquisition card 15 and transmitted to the host computer 7 for processing. In the circuit, the signal generator 13 provides signal drive for the quantum cascade laser 12 and the lock-in amplifier 14. The gas to be measured enters the mass flow meter 2 under the control of the first proportional valve 1. After flow control, it enters the measuring gas chamber 9 and circulates inside the measuring gas chamber 9. A pressure sensor 4 is installed at the outlet of the measuring gas chamber 9 to monitor the internal pressure of the measuring gas chamber 9 in real time. Then, it is connected to the vacuum pump 6 after passing through the second proportional valve 5 for pressure control. The pressure control unit 3 controls the system pressure at 30 Torr, with fluctuations within ±0.02 Torr. The temperature control unit 16 controls the system temperature at 22℃, with fluctuations within ±0.01℃. The pressure control unit 3 and the temperature control unit 16 transmit the collected pressure and temperature data to the host computer 7 in real time.

[0073] like Figure 2 As shown, an embodiment of the present invention provides a high-precision gas detection adaptive light intensity noise suppression method, based on the above-described system, comprising the following steps:

[0074] Step 1: A reference gas cell is placed in series in the optical path between the light source and the measurement gas cell. The reference gas cell adopts a wedge-shaped structure design to avoid interference noise from the parallel plates and to prevent some feedback light from entering the laser. After passing through the reference gas cell and the measurement gas cell, the laser beam is received by the detector. The reference gas has no cross-interference with the measurement gas in the absorption spectrum.

[0075] According to the Lambert-Beer law, when using TDLAS technology based on coaxial sensing for gas detection, the relationship between incident light intensity and transmitted light intensity can be expressed as:

[0076] (1);

[0077] in, It is the initial light intensity. It measures light intensity. It is a natural exponential function, representing the exponential decay of light intensity. It refers to the intensity of the reference gas absorption line. It is a reference gas absorption line shape. It is the reference gas pressure. It is a reference gas concentration. It is the reference gas absorption optical path. It measures the intensity of gas absorption lines. It measures the shape of the gas absorption line. It measures gas pressure. It measures gas concentration. This measures the optical path length of gas absorption. As can be seen from the formula above, laser intensity noise is multiplicatively coupled with gas concentration, which may mask gas concentration information.

[0078] In practice, due to the presence of laser noise, the model for laser intensity is:

[0079] (2);

[0080] in, The initial light intensity under ideal conditions. This represents the intensity noise of the laser beam, which is a non-stationary random process that follows a certain order. distributed, The distribution function, This represents a time-varying mean function, indicating the drift of the laser power. This represents a time-varying variance function, indicating the rapid fluctuations in laser power.

[0081] Therefore, within one scan cycle, the light intensities received by the detector at the reference time and the measurement time are respectively:

[0082] (3);

[0083] in, To reference the laser scanning time corresponding to the peak absorption of the gas, To measure the laser scanning time corresponding to the gas absorption peak, To reference the initial laser intensity corresponding to the peak absorption of the gas, To measure the initial laser intensity corresponding to the gas absorption peak, since this change occurs within the same scan cycle, its impact is considered negligible. The non-stationary change rate of noise is relatively slow, and it can be assumed that the reference gas cell can synchronously sense the laser intensity noise. The second harmonic signal of each absorption peak is extracted using a lock-in amplifier:

[0084] (4);

[0085] in, The amplitude of the second harmonic signal of the reference gas; To measure the amplitude of the second harmonic signal of the gas; The system gain for the reference gas; To measure the system gain of the gas.

[0086] Step 2: Define the system state vector and observation vector, and establish the state transition equation and measurement equation;

[0087] The system state vector and observation vector are defined as follows:

[0088] (5);

[0089] in, This is the system state vector; This represents the system observation vector; express The gas concentration is measured continuously and expressed in units of voltage. This is the light intensity drift coefficient. ; for laser light intensity at any given time; Nominal light intensity; The measured value is the second harmonic amplitude of the gas. The measured reference gas second harmonic amplitude.

[0090] Establish the following state transition equations and measurement equations:

[0091] (6);

[0092] in, Represents the state transition matrix; Represents the observation matrix; This is process noise; For measuring noise.

[0093] Step 3: Construct a dual-timescale noise feature extraction strategy to decompose the perceived light intensity noise signal into drift features and fluctuation features.

[0094] To extract noise fluctuations, the residual of the reference signal is calculated using the reference air cell signal to adaptively adjust the noise matrix. The estimated measurement noise variance is expressed as:

[0095] (7);

[0096] (8);

[0097] in, This represents the estimated variance of the measurement noise; It is the average amplitude of the second harmonic of the reference gas within the window; Indicates the scaling factor; This represents the size of the sliding window used to estimate the variance; This represents the forgetting factor.

[0098] To extract the drift trend, a weighted moving average method was used to eliminate the influence of short-term laser intensity fluctuations. The trend characteristics of the reference signal can be expressed as:

[0099] (9);

[0100] in, Indicates the laser drift trend; This represents the sliding window size for drift trend extraction; the observation matrix is... Measurement noise covariance matrix The update is as follows:

[0101] .

[0102] Step 4: Construct an adaptive estimation model;

[0103] Predicting system state and error covariance:

[0104] (10);

[0105] (11);

[0106] in, The predicted value of the state vector; This represents the predicted value of the error covariance matrix at the current time. This represents the predicted value of the error covariance matrix at the previous time step; Represents the process noise covariance matrix; state transition matrix ,in This is the gas concentration diffusion correction coefficient; a monitoring-reset strategy is adopted to enhance the tracking response capability of the algorithm. This strategy comprehensively considers current and past information data to determine whether to reset the estimated state; the decision criteria are set as follows:

[0107] (12);

[0108] (13);

[0109] in, Represents the innovation vector. Indicates the monitoring threshold. Indicates the length of the historical information sequence.

[0110] Calculate the gain matrix:

[0111] (14);

[0112] in, Represents the observation matrix. This represents the predicted value of the error covariance matrix;

[0113] Status Update:

[0114] (15);

[0115] Error covariance update:

[0116] (16);

[0117] in, This represents the predicted value of the error covariance matrix at the next time step. To robustly adjust the coefficients, It is an identity matrix.

[0118] Long-term testing was conducted on standard methane gas at a concentration of 300 ppmv, with continuous sampling for 8000 seconds. An adaptive light intensity noise suppression method was used simultaneously in real-time to process the raw data. The raw data and the data obtained using the adaptive light intensity noise suppression method are shown below. Figure 3 As shown in the figure, long-term measurement experimental data show that the adaptive light intensity noise suppression method improves the long-term stability of the detection system by 7.4 times. Allen bias calculations were performed on the steady-state data (2500~8000 s), and the results are as follows: Figure 4 As shown, under the conditions of a short optical path of 5 meters and an integration time of 0.5 seconds, the system can achieve a detection precision of 0.017 ppmv, which is more than 100 times higher than the original signal, effectively supporting continuous and reliable monitoring.

[0119] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high-precision gas detection adaptive light intensity noise suppression method, characterized in that, The method comprises the following steps: Step 1: driving the quantum cascade laser to emit laser by a signal generator, the laser is received by a detector after sequentially passing through a reference gas chamber and a measurement gas chamber, and the second harmonic signal amplitudes of the reference gas and the measurement gas are collected by a data acquisition card after demodulation by a lock-in amplifier and are transmitted to an upper computer; Step 2: defining a system state vector and an observation vector, establishing a state transition equation and a measurement equation; Step 3: constructing a double-time-scale noise feature extraction strategy, decomposing the perceived light intensity noise signal into drift features and fluctuation features; Step 4: constructing an adaptive estimation model, predicting and updating the system state and error covariance, and using a monitoring-reset strategy to enhance the tracking response capability, and finally outputting the measurement gas concentration value with the light intensity noise suppressed; The step 2 comprises the following specific steps: The system state vector and the observation vector are defined as: （5）； wherein, is the system state vector; is the system observation vector; denotes is the measured gas concentration at time t, expressed in voltage units; is the light intensity drift coefficient, ; is the is the laser light intensity at time t; is the nominal light intensity; is the measured measured gas second harmonic amplitude, is the measured reference gas second harmonic amplitude; The following state transition equation and measurement equation are established: （6）； wherein, denotes a state transition matrix; denotes an observation matrix; is a process noise; is a measurement noise; In the step 3, to extract noise fluctuation, the residual of the reference signal is calculated using the reference gas chamber signal to adaptively adjust the noise matrix; the estimated measurement noise variance is represented as: （7）； （8）； wherein, represents the estimated measurement noise variance; is the average amplitude of the reference gas second harmonic within the window; represents a scaling factor; represents the size of the sliding window for the estimation variance; represents a forgetting factor; To extract the drift trend, a weighted moving average method is used to eliminate the influence of short-term laser intensity fluctuations; the trend feature of the reference signal is represented as: （9）； wherein, represents a laser drift trend; represents a sliding window size for drift trend extraction; the observation matrix is ; the measurement noise covariance matrix is updated as: 。 2. The high-precision gas detection adaptive light intensity noise suppression method according to claim 1, characterized in that, In the step 1, according to the Lambert-Beer law, when using TDLAS technology based on co-axial sensing for gas detection, the relationship between the incident light intensity and the transmitted light intensity is represented as: （1）； wherein, is the initial light intensity, is the measured light intensity, is a natural exponential function representing the exponential decay of the light intensity, is the reference gas absorption line intensity, is the reference gas absorption line line shape, is the reference gas pressure, is the reference gas concentration, is the reference gas absorption optical path, is the measured gas absorption line intensity, is the measured gas absorption line line shape, is the measured gas pressure, is the measured gas concentration, is the measured gas absorption optical path; In practice, due to the existence of laser noise, the model of laser intensity is as follows: （2）； where, is the initial optical intensity in the ideal case; represents the optical intensity noise of the laser, which is a non-stationary random process, obeying distribution; is the distribution function; represents a time-varying mean function, indicating the drift of the laser power; represents a time-varying variance function, indicating the rapid fluctuations of the laser power; Therefore, within one scanning period, the light intensities received by the detector at the reference time and the measurement time are respectively: （3）； wherein, is the laser scanning time corresponding to the reference gas absorption peak, is the laser scanning time corresponding to the measurement gas absorption peak, is the initial laser light intensity corresponding to the reference gas absorption peak, is the initial laser light intensity corresponding to the measurement gas absorption peak; the second harmonic signal of each absorption peak is extracted using a lock-in amplifier as: （4）； wherein, is the reference gas second harmonic signal amplitude; is the measured gas second harmonic signal amplitude; is the system gain for the reference gas; is the system gain for the measured gas.

3. The high-precision gas detection adaptive optical intensity noise suppression method according to claim 2, characterized in that, In the step 4, the system state and error covariance are predicted: （10）； （11）； wherein is the predicted value of the state vector; represents the predicted value of the error covariance matrix at the current time instant; represents the predicted value of the error covariance matrix at the previous time instant; represents the process noise covariance matrix; state transition matrix wherein is the gas concentration diffusion correction coefficient; the tracking response capability of the algorithm is enhanced by using a monitoring-reset strategy, which considers the current and past innovation data to decide whether to reset the estimated state; the decision criteria are set as follows: （12）； （13）； wherein, denotes an innovation vector, denotes a monitoring threshold, denotes a length of the historical innovation sequence; The gain matrix is calculated: （14）； wherein, denotes an observation matrix, denotes a prediction of the error covariance matrix; State update: （15）； Error covariance update: （16）； wherein denotes a prediction of the error covariance matrix at the next time instant, is a robustness adjustment coefficient, is the identity matrix.

Images