Saturation absorption concentration detection method of single absorption peak TDLAS-WMS system

Abstract

This invention discloses a saturated absorption line concentration detection method for a single-absorption peak TDLAS-WMS system. This method, which belongs to the field of infrared laser absorption spectroscopy, first measures the characteristic values ​​of the absorption line of the second harmonic waveform under saturation. A fitting model is then established using a PSO-BPNN neural network. Multiple sets of concentration characteristic value data are used to train the model. Finally, the characteristic values ​​of the second harmonic waveform are measured to measure gas concentrations over a wide range. Because this method only requires scanning a single absorption peak, it is universally applicable and can achieve a fast response, shortening the system's response time when the measured gas concentration changes.

Description

Technical Field

[0001] The present invention belongs to the field of infrared laser absorption spectroscopy, and in particular relates to a saturation absorption concentration detection method applicable to a single absorption peak TDLAS-WMS system. Background Art

[0002] The importance of oil and gas to today's economic development is self-evident. As a non-renewable energy source, any discovered oil and gas reservoir faces resource depletion. Continuous discovery of new oil and gas reservoirs is crucial to meeting the sustained demand for oil and gas resources in today's rapidly developing society. The gaseous components of oil and gas primarily consist of hydrocarbon gases, including methane, ethane, and propane, as well as non-hydrocarbon compounds such as carbon dioxide, hydrogen sulfide, and nitrogen. Methane, a key component of oil and gas, makes monitoring its content a crucial tool for analyzing its genetic classification. Generally speaking, during oil and gas drilling and production, methane concentrations fluctuate continuously with changes in formation structure and drilling depth, typically ranging from a few hundred ppmv to tens of thousands of ppmv. Therefore, achieving wide dynamic range and high sensitivity methane concentration measurement is crucial for analyzing its genetic classification.

[0003] At present, gas detection systems based on absorption spectroscopy technology have the potential to perform on-site concentration detection due to their advantages such as high sensitivity, high stability, high accuracy, and rapid response, and have been successfully applied in many fields, such as atmospheric environment monitoring, pipeline leak warning, earthquake warning, deep-sea resource exploration, medical detection and analysis, etc. However, due to limitations such as optical power, optical path length, and absorption line intensity, the detection range of the absorption spectroscopy system is limited, and it is impossible to achieve accurate concentration measurement under a wide dynamic measurement range. Researchers have proposed a variety of effective methods to improve the detection dynamic range of the absorption spectroscopy system. S.ZHANG dynamically adjusts the laser output light intensity by adjusting the laser output light power, the number of light source flashes, and the integration time to ensure that the remaining light intensity after saturation can be received by the photodetector, thereby achieving chemical oxygen demand measurement under a wide dynamic range. However, this method of increasing light intensity by changing laser parameters can cause system signal drift, resulting in reduced detection performance. Another method is to select gas cells with different effective optical path lengths to achieve measurements at different concentrations. The higher the concentration of the gas to be measured, the shorter the effective optical path length of the selected gas cell. For example, Trios Company selected gas cells with different optical path lengths of 2mm, 5mm, and 20mm, and achieved a wide dynamic range of measurement from 1.5mg / L to 1100mg / L using a UV absorption spectroscopy system. ZhenWang combined WM-DAS and CRDS spectroscopy techniques to achieve both low detection limits and a wide measurement range, achieving carbon concentration measurements from 4ppmv to 10,000ppmv and a measurement accuracy of 0.5ppmv within the 4ppmv to 101ppmv concentration range. Both systems improve the dynamic range by changing the system structure. Different system structures are used to measure concentrations within different ranges. This approach complicates the system and may introduce additional interference noise. Zhiwei Liu achieved the measurement of carbon dioxide at different concentrations by selecting multiple absorption lines with different absorption intensities. He selected the absorption line with strong absorption intensity of 2315.20 cm -1 , and the weak absorption intensity of 2315.10 cm -1 Two carbon dioxide absorption lines were used to measure dissolved carbon dioxide in seawater, achieving a detection range of 11 ppbv to 1000 ppmv. However, this method can only be applied when there are lines of varying intensities near the selected line, making it not universally applicable. Summary of the Invention

[0004] The purpose of the present invention is to combine the measurement of second harmonic waveform parameters in the existing technology with a neural network to achieve regression prediction of concentration under saturated absorption conditions. A method for improving the gas concentration detection range of the TDLAS-WMS system is proposed. The characteristic values ​​of the absorption spectrum of the second harmonic waveform under saturation conditions are measured, and a fitting model is established through a PSO-BPNN neural network. The purpose of model training is achieved through multiple sets of concentration characteristic value data of the system. Finally, the characteristic values ​​of the second harmonic waveform are measured to achieve gas concentration measurement over a wide range.

[0005] The technical solution adopted by the present invention to achieve the above-mentioned object is: a single absorption peak TDLAS-WMS system saturation absorption concentration detection method, comprising the following steps, and the following steps are performed in sequence:

[0006] Step 1: Determine the central wavelength based on the absorption peak of the gas being measured. The microprocessor controls the driver and signal generator to generate a sawtooth wave, a single-frequency sine wave, and a co-directional double-frequency sine wave. The single-frequency sine wave and the sawtooth wave are provided to the laser through an adder to complete the scanning of the absorption peak band and sinusoidal modulation. The co-directional double-frequency sine wave is provided to the lock-in amplifier as a reference signal for extracting the second harmonic of the absorption peak signal.

[0007] Step 2: The microprocessor serves as the acquisition module of the TDLAS-WMS system to collect the maximum value point Vmax, the minimum value point Vmin of the second harmonic of the absorption peak, and the sampling point difference Δl between the maximum value point Vmax and the minimum value point Vmin of the second harmonic of the absorption peak, and transmits the maximum value point Vmax, the minimum value point Vmin of the second harmonic of the absorption peak, and the sampling point difference Δl between the maximum value point Vmax and the minimum value point Vmin of the second harmonic of the absorption peak to the host computer as three characteristic value points;

[0008] Step 3: Use the near-infrared analysis chamber to detect the standard methane gas concentration in order from low concentration to high concentration. The effective concentration data of each concentration detection cycle is recorded after multiple smoothing filters by the host computer. The three eigenvalues ​​of the maximum value point Vmax, minimum value point Vmin of the second harmonic of the absorption peak detected at each concentration, and the difference Δl in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the second harmonic are used as a set of neural network input data. The standard gas concentration data is used as the response data to train and verify the neural network. The root mean square error between the predicted value and the true value is used as the loss function for model training. Based on the PSO-BPNN neural network, a functional relationship model between the eigenvalue points and the measured gas concentration is established, and finally the nonlinear functional relationship between the three eigenvalue points and the measured gas concentration is quantified;

[0009] Step 4: Bring the functional relationship model between the characteristic value points and the measured gas concentration into the TDLAS-WMS system. By detecting the three characteristic values ​​of the second harmonic waveform (maximum point Vmax, minimum point Vmin, and the difference Δl in the number of sampling points between the maximum point Vmax and the minimum point Vmin of the second harmonic), the concentration in the saturated region can be calculated.

[0010] Furthermore, in step 1, a microprocessor controls the drive and signal generator to generate a laser scanning drive signal, a laser modulation signal, and a lock-in amplifier reference signal, and the laser scanning drive signal generates a trigger square wave signal through a comparator; then the laser scanning drive signal and the laser modulation signal generate a laser drive signal, which drives the laser of the TDLAS-WMS system to complete the scanning of the absorption peak band, and the lock-in amplifier reference signal is provided to the lock-in amplifier to realize the second harmonic extraction of the absorption peak signal.

[0011] Furthermore, the laser scanning driving signal covers a target absorption peak band, and the target absorption peak is any single absorption peak.

[0012] Furthermore, in step 2, the sampling point difference Δl between the maximum point Vmax and the minimum point Vmin of the second harmonic is the difference between the maximum point Vmax and the minimum point Vmin of the selected second harmonic waveform.

[0013] Furthermore, in step 3, the PSO-BPNN neural network uses the PSO algorithm to optimize the weight and threshold parameters in the network and uses the BPNN model as the network structure of the neural network; the loss function of the training model is as follows:

[0014] ReLU(x)=max(0,x) (1)

[0015]

[0016] Among them, ReLU is the activation function, max is the maximum function, Loss function is the loss function, RMSE is the root mean square error function, f is the model function, x is the model input, y is the model output, i is the number of data groups, and N is the total number of data groups.

[0017] Furthermore, the complex nonlinear functional relationship between the three characteristic value points and the measured gas concentration is as follows: when the gas concentration reaches 1200ppmv, the measured gas produces strong absorption, the direct absorption waveform shows a clipping phenomenon, and the second harmonic waveform is distorted. This concentration region is the saturation region; before the saturation phenomenon occurs, the characteristic value points and the concentration are linearly related. After the saturation phenomenon occurs, as the gas concentration increases, within different concentration ranges, the maximum point Vmax and the concentration, and the minimum point Vmin and the concentration show a certain nonlinear functional relationship.

[0018] The above-mentioned design scheme can bring the following beneficial effects: The present invention proposes a saturated absorption concentration detection method for a single-absorption peak TDLAS-WMS system. This method first measures the characteristic values ​​of the absorption spectrum of the second harmonic waveform under saturation, then establishes a fitting model using a PSO-BPNN neural network. This model is trained using multiple sets of concentration characteristic value data from the system, and finally, gas concentration measurements over a wide range are achieved by measuring the characteristic values ​​of the second harmonic waveform. Because this method only requires scanning a single absorption peak, it is universally applicable and can achieve rapid response, shortening the system's response time when the measured gas concentration changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to understand the present invention and do not constitute improper limitations of the present invention. In the drawings:

[0020] Figure 1 Second harmonic simulation diagrams of different regions of the single absorption peak TDLAS-WMS system saturation absorption concentration detection method proposed in the present invention;

[0021] Figure 2 This is a structural diagram of the device used in the saturation absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention;

[0022] Figure 3 This is a detection flow chart of the saturation absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention;

[0023] Figure 4 This is a flowchart of the neural network modeling in the saturation absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention;

[0024] Figure 5 This is a neural network model diagram for the saturation absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention;

[0025] Figure 6 This is a diagram showing the oil and gas methane concentration measurement results in the saturated absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention;

[0026] Figure 7 This is the Allan variance diagram of the measurement data of the saturation absorption concentration detection method of the single absorption peak TDLAS-WMS system proposed in the present invention. DETAILED DESCRIPTION

[0027] To more clearly illustrate the present invention, the present invention is further described below with reference to preferred embodiments and the accompanying drawings. Those skilled in the art should understand that the following detailed description is illustrative rather than restrictive and should not be used to limit the scope of protection of the present invention. Unless otherwise defined, technical or scientific terms used herein should have the same meaning as those having ordinary skill in the art to which the present invention belongs.

[0028] The present invention provides a single absorption peak TDLAS-WMS system saturation absorption concentration detection method, which is suitable for wide-range concentration detection of lasers in the TDLAS-WMS system. The method firstly controls the drive and signal generator by a microprocessor to generate a laser scanning drive signal, a laser modulation signal and a lock-in amplifier reference signal, and the laser scanning drive signal generates a trigger square wave signal through a comparator; then the laser scanning drive signal and the laser modulation signal generate a laser drive signal, which drives the laser of the TDLAS-WMS system to complete the scanning of the absorption peak band, and the lock-in amplifier reference signal is provided to the lock-in amplifier to realize the second harmonic extraction of the absorption peak signal, and then the microprocessor is used as an acquisition module of the TDLAS-WMS system to acquire the maximum value point Vmax and the minimum value point Vmin of the second harmonic signal of the absorption peak, and the second harmonic The sampling point difference Δl between the maximum point Vmax and the minimum point Vmin of the second harmonic is calculated, and the sampling point difference Δl between the maximum point Vmax and the minimum point Vmin of the second harmonic is the difference between the maximum point Vmax and the minimum point Vmin of the selected second harmonic waveform, and is transmitted to the host computer; the standard methane gas concentration is detected in order from low concentration to high concentration using a near-infrared analysis cavity, and the three eigenvalues ​​of the maximum point Vmax, the minimum point Vmin and the sampling point difference Δl between the maximum point Vmax and the minimum point Vmin of the second harmonic obtained at each concentration are used as a set of neural network input data, and the standard gas concentration data is used as response data to train and verify the PSO-BPNN neural network, and then a functional relationship model between the eigenvalue points and the measured gas concentration is established based on the PSO-BPNN neural network, such as Figure 5 As shown in the figure, the three eigenvalues ​​of the maximum value point Vmax, the minimum value point Vmin, and the difference Δl in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the second harmonic signal of the absorption peak detected at different concentrations are used as a set of neural network input data, and the standard gas concentration data is used as the response data to train and verify the network. Finally, the functional relationship model between the eigenvalue points and the measured gas concentration established based on the PSO-BPNN neural network is brought into the TDLAS-WMS system. By detecting the three eigenvalues ​​in the second harmonic waveform, the concentration calculation in the saturation region is realized.

[0029] The TDLAS-WMS system is composed of a laser, an optical cavity, a detector, a main control circuit and related control modules. The TDLAS-WMS system belongs to the prior art. In order to avoid confusing the essence of the present invention, the known components and circuits are not described in detail.

[0030] In order to better understand the technical solution of the present invention, the present invention will be further described below in combination with the principle calculation of the present invention and the accompanying drawings of the specification;

[0031] Infrared absorption spectroscopy is used to measure gas concentration by analyzing the absorption intensity of the absorption peak in the infrared region. This process is quantitatively described by the Lambert-Beer law:

[0032] I(v)=I0e -α(v) (3)

[0033]

[0034] Where I(v) is the outgoing light intensity, I0 is the initial light intensity, α(v) is the gas absorption coefficient at the beam frequency v, and S is the gas absorption line intensity at the beam frequency v, in cm -2 / atm, P is the pressure of the gas in the optical cavity, C is the concentration of the gas to be measured, L is the effective optical path of the light beam transmission, is the absorption line shape of the gas.

[0035] In the TDLAS-WMS system, a high-frequency sinusoidal modulation signal is added to the laser's output light, as shown in Equation (5). The amplitude of the second harmonic signal is then detected to measure the concentration of the gas being measured. This method shifts the system's operating frequency band to the high-frequency region, effectively preventing the impact of low-frequency drift on the system's detection performance.

[0036] v=v c +m·cos(wt) (5)

[0037] where v c is the center frequency of the gas absorption peak, m is the laser wavelength modulation amplitude, w is the angular frequency of the laser modulation signal, and t is time. The absorption function is expanded in the cosine Fourier series:

[0038]

[0039] Among them A n is the Fourier series of the nth harmonic. In order to characterize the relationship between the output beam intensity and the gas concentration, the measured concentration is divided into three regions from small to large: linear region, nonlinear region and saturation region.

[0040] When the absorption coefficient is small, α(v)<<1, and equation (3) can be simplified to:

[0041]

[0042] When the incident light frequency is equal to the center frequency v of the gas absorption peak c When , the absorption peak second harmonic Fourier coefficient A2 reaches its maximum value, and formula (7) is simplified to:

[0043] I 2fmax =-kI0SCLA2 (8)

[0044] Where k is a constant affected by the system modulation depth. Equation (8) shows a linear relationship between the measured gas concentration and the second harmonic peak. The relationship between the outgoing light intensity and the concentration C is approximately linear, i.e., in the linear region. When the absorption coefficient is large, the Lambert-Beer law cannot be approximated by Equation (7). The relationship between the outgoing light intensity and the concentration C is a nonlinear relationship with an exponential e, as shown in Equation (9). The second harmonic peak also increases nonlinearly with increasing concentration, i.e., in the nonlinear region.

[0045]

[0046] When the gas concentration increases further, the measured gas will produce strong absorption. In this case, the output light intensity is close to 0, the direct absorption waveform will be clipped, and the second harmonic waveform will also be distorted. This concentration area is the saturation area. The second harmonic simulation diagrams in different areas are as follows: Figure 1 shown.

[0047] like Figure 2 As shown in the figure, a device used in a single absorption peak TDLAS-WMS system saturation absorption concentration detection method is divided into three parts: gas part, optical part and circuit part. The microprocessor controls the drive and signal generator to generate a drive signal that is transmitted to the laser to excite the laser to output detection light of the corresponding wavelength of the spectrum line to be measured, which is then injected into the gas chamber to measure the measured gas; the detector obtains the direct absorption spectrum line and provides it together with the same direction double frequency sine wave to the phase-locked amplifier to extract the second harmonic of the absorption peak and transmit it to the host computer; the host computer processes the measurement data according to the established functional relationship between the characteristic points and the gas concentration, and then realizes the calculation of the gas concentration in the saturation region. The specific detection process is as follows: Figure 3 shown.

[0048] In order to quantify the complex nonlinear functional relationship between the three eigenvalue points and the measured gas concentration, a functional relationship model between the eigenvalue points and the measured gas concentration is established based on the PSO-BPNN neural network. The error back propagation algorithm in the BPNN model has the problems of slow iteration speed and large error. In order to improve the fitting accuracy and iteration speed of the BPNN model, the PSO algorithm is used to optimize the weight and threshold parameters in the BPNN model. Finally, a functional relationship model between the eigenvalue points and the measured gas concentration is established based on the PSO-BPNN neural network. The modeling flow chart is as follows: Figure 4 shown.

[0049] First, a BPNN model was established. The BPNN model consists of an input layer, three hidden layers, and an output layer. The layer sizes of the three hidden layers are 20, 30, and 20, respectively. The ReLU function (Equation (1)) was used as the activation function of the BPNN model to ensure the training effect of the BPNN model. The second harmonic maximum point Vmax, minimum point Vmin, and the difference in the number of sampling points Δl between the maximum point Vmax and the minimum point Vmin of the second harmonic measured by the system under multiple sets of standard gas concentrations were used as the input values ​​of the training set data. The measured standard gas concentration was used as the response value of the BPNN model, and the root mean square error between the predicted value and the true value was used as the loss function for model training, as shown in Equation (2).

[0050] After determining the BPNN model, the initial population of the PSO algorithm is established. The population size is 50 and the maximum number of iterations is 100. Each particle in the population includes the weight w of the hidden layer in the BPNN model. i and threshold b i The information is collected and a random number between -1 and 1 is used as the initial value of the particle parameter. The parameter optimization of the BPNN model is completed by continuously iterating and optimizing the population particle parameters.

[0051] Taking methane concentration measurement as an example, the implementation process of the single absorption peak TDLAS-WMS system saturation absorption concentration detection method proposed in the present invention includes the following steps:

[0052] Step 1. To ensure that the selected absorption spectral lines can be separated during the measurement process and to avoid the influence of spectral line overlap on the measurement results, the pressure of the gas chamber is controlled: the gas chamber pressure is controlled at 760 Torr, and the pressure fluctuation is within 0.2 Torr; the microprocessor controls the driver and signal generator to generate a sawtooth wave with a frequency of 3 Hz, and simultaneously generates two in-phase sine waves with frequencies of 3 kHz and 6 kHz, respectively. The sine wave with a frequency of 3 kHz and the sawtooth wave are provided to the laser after passing through the adder to complete the scanning and sine modulation of the absorption peak band. The sine wave with a frequency of 6 kHz is provided to the lock-in amplifier as the reference signal to realize the second harmonic extraction of the absorption peak signal;

[0053] Step 2: The microprocessor serves as the acquisition module of the TDLAS-WMS system, and collects the maximum value point Vmax, the minimum value point Vmin of the absorption peak second harmonic, and the sampling point difference Δl between the maximum value point Vmax and the minimum value point Vmin of the second harmonic as characteristic value points, and transmits them to the host computer;

[0054] Step 3: To obtain the training, verification, and test data required for the PSO-BPNN neural network, standard methane gas and nitrogen are mixed with standard methane gas of different concentrations (100 ppmv to 14,000 ppmv) through the gas distribution system. The standard methane gas concentration is detected in the gas chamber from low concentration to high concentration. After ensuring that the gas in the gas chamber is fully purged, the effective concentration data of each concentration detection cycle is recorded after multiple smoothing filters by the host computer. The measured concentration group includes three parts: the linear region, the nonlinear region, and the saturation absorption region of the second harmonic peak. The three characteristic values ​​of the second harmonic maximum point Vmax, the minimum point Vmin, and the difference Δl in the number of sampling points between the maximum point Vmax and the minimum point Vmin of the second harmonic obtained at each concentration are used as a set of neural network input data, and the standard gas concentration data is used as the response data to train and verify the network.

[0055] Step 4: By establishing a functional relationship model between the eigenvalue points and the measured gas concentration based on the PSO-BPNN neural network, a complex nonlinear functional relationship between the three eigenvalue points and the measured gas concentration is obtained. A long-term measurement of 1400s is performed on the standard gas with a concentration of 300ppmv, and the detection limit performance of the system is characterized by the Allan variance. Figure 6 and Figure 7 As shown in the figure, processing and analyzing the raw data collected by the TDLAS-WMS system revealed an Allan variance plot calculated from the concentration data, demonstrating a detection limit of 0.251 ppmv at an integration time of 1 second, and a minimum detection limit of 0.111 ppmv at an integration time of 54 seconds. The TDLAS-WMS system based on this method can measure concentrations over a wide range of 0.111 to 14,000 ppmv.

[0056] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. Single absorption peak TDLAS-WMS system saturation absorption concentration detection method, characterized in that: The process includes the following steps, which are performed in sequence: Step 1: Determine the central wavelength based on the absorption peak of the gas being measured. The microprocessor controls the driver and signal generator to generate a sawtooth wave, a single-frequency sine wave, and a co-directional double-frequency sine wave. The single-frequency sine wave and the sawtooth wave are provided to the laser through an adder to complete the scanning of the absorption peak band and sinusoidal modulation. The co-directional double-frequency sine wave is provided to the lock-in amplifier as a reference signal for extracting the second harmonic of the absorption peak signal. Step 2: The microprocessor is used as the acquisition module of the TDLAS-WMS system to collect the maximum value point Vmax, the minimum value point Vmin of the second harmonic of the absorption peak, and the difference between the maximum value point Vmax and the minimum value point Vmin of the second harmonic of the absorption peak. The acquisition is performed, and the maximum value point Vmax, the minimum value point Vmin of the absorption peak second harmonic and the difference in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the absorption peak second harmonic are calculated. Transmitted to the host computer as three characteristic value points; Step 3: Use the near-infrared analysis chamber to detect the standard methane gas concentration in the order from low concentration to high concentration. The effective concentration data of each concentration detection cycle is recorded after multiple smoothing filters by the host computer. The maximum value point Vmax, minimum value point Vmin of the second harmonic of the absorption peak detected at each concentration, and the difference in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the second harmonic are recorded. The three eigenvalues ​​are used as a set of neural network input data, and the standard gas concentration data is used as the response data to train and verify the neural network. The root mean square error between the predicted value and the true value is used as the loss function for model training. Based on the PSO-BPNN neural network, a functional relationship model between the eigenvalue points and the measured gas concentration is established, and finally the nonlinear functional relationship between the three eigenvalue points and the measured gas concentration is quantified. Step 4: Bring the functional relationship model between the characteristic value point and the measured gas concentration into the TDLAS-WMS system, and detect the maximum value point Vmax, minimum value point Vmin and the difference in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the second harmonic waveform. These three eigenvalues ​​can realize the concentration calculation in the saturated region; In step 3, the PSO-BPNN neural network uses the PSO algorithm to optimize the weight and threshold parameters in the network and uses the BPNN model as the network structure of the neural network; the loss function of the model training is as follows: （1） （2） Where ReLU is the activation function, max is the maximum function, Loss function is the loss function, RMSE is the root mean square error function, f is the model function, x is the model input, y is the model output, i is the number of data sets, and N is the total number of data sets; The complex nonlinear functional relationship between the three characteristic value points and the measured gas concentration is as follows: when the gas concentration reaches 1200ppmv, the measured gas produces strong absorption, the emitted light intensity is close to 0, the direct absorption waveform shows a clipping phenomenon, and the second harmonic waveform is distorted. The area of ​​this concentration is the saturation area; before the saturation phenomenon occurs, the characteristic value points and the concentration are linearly related. After the saturation phenomenon occurs, as the gas concentration increases, within different concentration ranges, the maximum point Vmax and the concentration, and the minimum point Vmin and the concentration show a certain nonlinear functional relationship.

2. The single absorption peak TDLAS-WMS system saturation absorption concentration detection method according to claim 1, characterized in that: In step 1, the microprocessor controls the driver and signal generator to generate a laser scanning drive signal, a laser modulation signal, and a lock-in amplifier reference signal. The laser scanning drive signal generates a trigger square wave signal through a comparator. Then, the laser scanning drive signal and the laser modulation signal generate a laser drive signal to drive the laser of the TDLAS-WMS system to complete the scanning of the absorption peak band. The lock-in amplifier reference signal is provided to the lock-in amplifier to realize the second harmonic extraction of the absorption peak signal.

3. The single absorption peak TDLAS-WMS system saturation absorption concentration detection method according to claim 2, characterized in that: The laser scanning driving signal covers a target absorption peak band, and the target absorption peak is any single absorption peak.

4. The single absorption peak TDLAS-WMS system saturation absorption concentration detection method according to claim 1, characterized in that: In step 2, the difference in the number of sampling points between the maximum value point Vmax and the minimum value point Vmin of the second harmonic is It is the difference between the maximum value point Vmax and the minimum value point Vmin of the selected second harmonic waveform.

Images