A TDLAS Combined Detection Method

Abstract

The application provides a TDLAS combined detection method, which is based on an existing gas two-dimensional concentration distribution reconstruction method based on TDLAS remote detection, and combines a TDLAS in-situ detection system and a TDLAS remote detection system to detect indoor gas concentration, optimizes the gas detection method, corrects the iteration result by using in-situ detection data in each iteration of the reconstruction process, improves the image reconstruction algorithm, and proposes a gas two-dimensional concentration distribution reconstruction method based on TDLAS combined detection. The application improves the reconstruction accuracy and reduces the reconstruction time.

Description

Technical Field

[0001] This invention belongs to the field of gas detection, specifically relating to a TDLAS combined detection method. Background Technology

[0002] Most people spend 90% of their time indoors, making the indoor gas environment crucial. Gas Distribution Mapping (GDM) technology can reveal the characteristics of gas diffusion in the environment. Currently, most methods for establishing GDMs rely on contact gas sensors or sensor networks. However, this point-based measurement method, due to the sparse measurement points, cannot establish the concentration distribution across the entire space. To achieve higher spatial coverage, linear gas detection technology is needed to establish indoor GDMs. Tunable Diode Laser Absorption Spectroscopy (TDLAS) is an effective gas detection technology. It obtains the concentration of the analyte by detecting the absorption intensity of absorption lines. It offers advantages such as non-contact operation, high sensitivity, high accuracy, high selectivity, rapid measurement, and no pre-processing required, and has been widely used in various fields. TDLAS technology can be divided into in-situ detection and telemetry based on the measurement mode. In-situ detection involves directly measuring the concentration of the analyte gas in the gas cell within the target area, treating the measurement range approximately as a single point, resulting in higher detection accuracy than telemetry. TDLAS telemetry technology can measure the average gas concentration along a laser path over a wide open optical path, with telemetry distances ranging from tens to hundreds of meters, and the measurement range can be approximated as a line. Combining TDLAS with in-situ detection and telemetry technologies can achieve higher spatial coverage and detection accuracy.

[0003] Research on reconstructing indoor gas concentration distribution (GDM) using TDLAS combined detection technology has not yet emerged; current researchers only utilize TDLAS telemetry technology to reconstruct the two-dimensional gas concentration distribution. Reconstructing indoor GDM is an extended application of reconstructing the two-dimensional gas concentration distribution. The technical approach for reconstructing the two-dimensional gas concentration distribution based on TDLAS telemetry technology involves obtaining spectral data along a laser line using a TDLAS telemetry system, and then obtaining the two-dimensional gas concentration distribution map through Algebraic Reconstruction Technique (ART). The two-dimensional gas concentration distribution reconstruction technology based on TDLAS telemetry is widely used for the two-dimensional reconstruction of gas concentrations in flame flow fields and is relatively mature, such as for detecting CO, CO2, O2, and NO during combustion. XThe two-dimensional concentration distribution of gases can provide timely information on the combustion status of the furnace, thereby enabling combustion control. However, the two-dimensional gas concentration distribution reconstruction technology based on TDLAS telemetry still has room for improvement in reconstruction accuracy, and the reconstruction time is relatively long, limiting its application in other fields.

[0004] In traditional TDLAS telemetry-based two-dimensional gas concentration distribution reconstruction techniques for flame flow fields, multiple laser transceivers are required to obtain different measurement data. This approach lacks the ability to explore unknown spaces and is complex. Mobile TDLAS technology, by controlling the detection position of the TDLAS telemetry system via a mobile platform, can acquire a large amount of measurement data. However, in the process of reconstructing indoor gas concentration distribution (GDM) using mobile TDLAS technology, detection points are randomly distributed within or along the grid lines of the reconstruction area. Existing TDLAS two-dimensional gas concentration distribution reconstruction techniques cannot be directly applied to mobile TDLAS technology for indoor GDM reconstruction. Therefore, there is an urgent need to research a detection method suitable for mobile TDLAS combined detection systems to improve reconstruction accuracy and reduce reconstruction time.

[0005] In summary, detecting the gas concentration distribution in unknown indoor spaces and establishing indoor gas concentration distribution mapping (GDM) is crucial. Most current research employs point-based measurement methods, resulting in sparse measurement points. Furthermore, traditional TDLAS (Transient Digital Leveraging) gas concentration distribution detection technology is based solely on TDLAS telemetry, leading to low reconstruction accuracy and slow reconstruction speed. Summary of the Invention

[0006] To address the issues of low reconstruction accuracy and long reconstruction time in reconstructing indoor concentration distribution maps using current TDLAS telemetry technology, this invention provides a TDLAS joint detection method. This method improves reconstruction accuracy and speed while allowing for flexible application in different scenarios and reducing detection costs. Based on a deep understanding of the principles of reconstructing two-dimensional gas concentration distribution using TDLAS telemetry technology, this invention, building upon existing TDLAS telemetry-based methods, combines a TDLAS in-situ detection system and a TDLAS telemetry system for indoor gas concentration detection. This optimizes the gas detection method by using in-situ detection data to correct iteration results in each iteration of the reconstruction process, thus improving the image reconstruction algorithm. This invention proposes a TDLAS joint detection-based method for reconstructing two-dimensional gas concentration distribution, which improves reconstruction accuracy and reduces reconstruction time.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A TDLAS combined detection method includes the following steps:

[0009] Step 1: Acquire TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment. Measure indoor gas concentration data using the TDLAS telemetry system and TDLAS in-situ detection system along the detection route.

[0010] Step 2: Reconstruct the two-dimensional concentration distribution map of indoor gases. Using TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment, the two-dimensional concentration distribution map of indoor gases is reconstructed according to the joint modified algebraic iterative algorithm.

[0011] Furthermore, in step 1,

[0012] The acquisition of indoor gas environment TDLAS telemetry data includes:

[0013] The TDLAS telemetry system is designed to detect gas concentrations along the detection route at a certain frequency.

[0014] Furthermore, in step 1, acquiring TDLAS in-situ detection data of the indoor gas environment includes:

[0015] While the TDLAS telemetry system detects gas concentration at a certain frequency, the TDLAS in-situ detection system detects gas concentration at the same frequency.

[0016] Further, in step 2, reconstructing the two-dimensional indoor gas concentration distribution map includes:

[0017] The two-dimensional concentration distribution map of indoor gases is reconstructed using a joint modified algebraic iterative algorithm, which includes the following steps:

[0018] Step (1) Reconstruct the two-dimensional distribution map of gas concentration using an algebraic iterative algorithm;

[0019] Step (2) employs an adaptive correction algebraic iterative algorithm to improve reconstruction accuracy and speed;

[0020] Step (3) uses non-negativity constraints and smoothness criteria to constrain the reconstruction process;

[0021] Step (4) Design a joint modified algebraic iterative algorithm to further improve reconstruction accuracy and reconstruction speed.

[0022] Further, in step (1), the image is first discretized, that is, the entire two-dimensional gas concentration distribution image f(x,y) is discretized into N grids, and the gas parameters within the grids are considered to be uniform; the gas concentration within the grids is obtained by inverting the gas absorption coefficient, and the grid value f i The gas absorption coefficient α in the i-th grid represents the gas absorption coefficient. i1≤n≤N, according to the Beer-Lambert law, when a laser beam j passes through a target region, the integral absorbance of the absorption line with center frequency v is discretely expressed as:

[0023]

[0024] Among them, A j Let S be the integrated absorbance of the j-th ray; S(T) is the linear intensity at temperature T.

[0025] The above equation is the gas absorption equation, where M is the total number of laser beams; [PXS(T)] i L is the product of pressure, concentration, and linear intensity at grid point i; i,j is the projection coefficient, which is numerically equal to the length of the laser beam j intercepted by the grid i;

[0026] In the actual detection process, detection lasers are emitted from different directions and angles to scan and reconstruct the area. Each laser beam has a gas absorption equation as described in equation (1). The gas absorption equations of each laser beam are combined and expressed as a set of linear equations as shown in equation (2):

[0027]

[0028] The reconstructed value of the gas absorption coefficient is obtained by solving the linear equation system. The gas absorption coefficient vector is set as α = (α1, α2, ..., α...). N The initial value vector of ) is the zero vector. Substitute it into equation (2) for iterative calculation, and then use the gas absorption coefficient vector α. (j-1) Substituting into the j-th equation of equation (2), we obtain α. (j) The iterative formula is:

[0029]

[0030] Where λ is the relaxation factor, 0 < λ < 2, which is related to the convergence speed and reconstruction quality; n represents the index of the N-dimensional image vector, 1 ≤ n ≤ N; according to the above iteration, α is obtained from the Mth equation of equation (2). (M) If the change in the gas absorption coefficient vector is greater than the threshold and the number of iterations is less than 10,000, then the gas concentration distribution can be solved by finding the linear intensity at room temperature in the Hisran database.

[0031] Further, step (2) includes:

[0032] An adaptive correction algebraic iterative algorithm is used to determine the gas absorption coefficient α within each grid cell. i The value is obtained by solving the following iterative formula:

[0033]

[0034]

[0035] j k =1+[k mod(M)] (4c)

[0036] Where k represents the number of iterations in the MAART iteration process; β is related to the adaptive adjustment step size and takes a value of 0.25; Let be the adaptive relaxation factor, representing the relaxation factor of grid i in the k-th iteration, and mod() be the modulo function.

[0037] Further, step (3) includes:

[0038] For the gas absorption coefficient, a non-negative constraint is added, and the non-negative constraint function Φ() is shown in equation (5):

[0039]

[0040] Meanwhile, a smoothing criterion is used to suppress abrupt changes in the reconstruction results. In each iteration, the gas absorption coefficient of grid i is smoothed using itself and 8 surrounding grids, a total of 9 points. The smoothing criterion is shown in equation (6):

[0041]

[0042] Where δ is the smoothing factor, and m and n in the above formula are the row and column indices of the absorption coefficient in the reconstructed region grid, respectively;

[0043] To eliminate the influence of pressure P and linear strength S(T) on the iteration results, a new weighting factor L' is defined. i,j :

[0044] L' i,j =P i ·L i,j ·S i (T i (7)

[0045] Among them, S i (T i ) represents the line strength of the i-th grid.

[0046] The integral absorbance at this point, also known as the projected value A, is expressed as:

[0047] A = X·L' (8)

[0048] Where L′ is the new weighting factor;

[0049] For each absorption line, if there are I projected rays, there will be I absorption equations. The linear equations about the gas absorption coefficient are transformed into a linear equations about the concentration X.

[0050] Further, step (4) includes:

[0051] For the gas absorption coefficient at the grid where the in-situ measurement is located, an in-situ correction is added in each iteration. The in-situ correction formula is as follows, where the correction factor χ = 0.01:

[0052]

[0053] In the above formula, α inSitu,i This represents the in-situ gas concentration measurement at grid i.

[0054] Beneficial effects:

[0055] (1) This invention proposes a two-dimensional gas concentration distribution reconstruction method based on TDLAS joint detection. Based on the traditional two-dimensional gas concentration distribution reconstruction method based on TDLAS telemetry, an in-situ detection method is added during the detection process. The reconstruction results are corrected by using in-situ detection data during the reconstruction process. Experiments have shown that this method can quickly and effectively reconstruct indoor gas distribution maps. In addition to reconstructing indoor gas distribution areas, it can also be used to reconstruct gas concentration distribution in enclosed places such as warehouses.

[0056] (2) Compared with the traditional gas two-dimensional concentration distribution reconstruction method based on TDLAS telemetry, the present invention has higher reconstruction speed and reconstruction accuracy. Attached Figure Description

[0057] Figure 1 This is a flowchart of the overall TDLAS joint detection method of the present invention;

[0058] Figure 2 This is a test roadmap for the TDLAS joint testing system;

[0059] Figure 3 A distribution map of light detected by the TDLAS telemetry system;

[0060] Figure 4 This is a distribution map of the detection points of the TDLAS in-situ detection system;

[0061] Figure 5 This is a diagram illustrating the algorithm improvement steps of the present invention;

[0062] Figure 6a , Figure 6b , Figure 6c , Figure 6d , Figure 6e , Figure 6f , Figure 6g , Figure 6h The images show reconstruction results based on TDLAS telemetry and reconstruction results based on TDLAS joint detection; among them, Figure 6a A 3D image of the reconstruction result from the ART algorithm. Figure 6b A top view of the reconstruction results from the ART algorithm. Figure 6c A stereoscopic image of the reconstruction result from the MAART algorithm. Figure 6d A top view of the reconstruction results from the MAART algorithm. Figure 6e To reconstruct the image using the MAART algorithm with non-negativity constraints and smoothing criteria, Figure 6f The reconstruction result of the MAART algorithm with non-negativity constraints and smoothing criteria added; Figure 6g To reconstruct the 3D image using the joint modified algebraic iterative algorithm, Figure 6h A top view of the reconstruction results of the joint modified algebraic iterative algorithm. Detailed Implementation

[0063] 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. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0064] This invention relates to the field of gas detection, and particularly to rapid gas detection for reconstructing two-dimensional gas concentration distributions. For example... Figure 1 As shown, the TDLAS combined detection method of the present invention includes the following steps:

[0065] Step 1: Acquire TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment. Measure indoor gas concentration data using the TDLAS telemetry system and TDLAS in-situ detection system along the predetermined detection route.

[0066] Step 2: Reconstruct the two-dimensional concentration distribution map of indoor gases. Using TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment, the two-dimensional concentration distribution map of indoor gases is reconstructed according to the joint modified algebraic iterative algorithm.

[0067] In step 1, acquiring TDLAS telemetry data of the indoor gas environment includes:

[0068] Before data acquisition, the detection route was designed. Taking a 20m×20m enclosed room as an example, the detection route was designed to be approximately square to obtain approximately orthogonally distributed laser beams. Simultaneously, the laser emission direction was rotated at low speed at the corners to obtain fan-shaped beams, improving reconstruction accuracy. The detection route is as follows: Figure 2 As shown in the figure, the starting point of the detection route is randomly selected near the wall of the room. The starting point is represented by the triangle in the upper left corner of the figure. The direction of travel is clockwise. During data acquisition, the detection light of the TDLAS telemetry system is perpendicular to the direction of travel, represented by the dashed line. The average speed during travel is about 1 m / s. The system stops at the corner and rotates only at an angular velocity of about 0.251 rad / s.

[0069] During the data acquisition process along the detection route by the TDLAS joint detection system, the TDLAS telemetry system within the TDLAS joint detection system detects gas concentration at a certain frequency. The detection range of the TDLAS telemetry system can be approximated as a line. Taking a detection interval of 1 second as an example, the distribution of the light rays detected by the TDLAS telemetry system during data acquisition is as follows: Figure 3 As shown, the dots represent the location of the telemetry system when the corresponding detection ray is generated, and the straight lines starting from the dots represent the detection rays.

[0070] In step 1, obtaining TDLAS in-situ detection data of the indoor gas environment includes:

[0071] While the TDLAS telemetry system detects gas concentration at a certain frequency, the TDLAS in-situ detection system within the TDLAS combined detection system also detects gas concentration at the same frequency. The detection range of the TDLAS in-situ detection system can be approximated as a single point. Taking a detection interval of 1 second as an example, the distribution of detection points of the TDLAS in-situ detection system during data acquisition is as follows: Figure 4 As shown, the dots represent the location of the detection system when the corresponding detection light is generated, which is also the detection point of the TDLAS in-situ detection system.

[0072] Step 2, reconstructing the two-dimensional indoor gas concentration distribution map, includes:

[0073] To obtain a two-dimensional distribution image of indoor gas concentration, TDLAS technology is combined with an image reconstruction algorithm. Currently, most TDLAS telemetry-based two-dimensional gas concentration distribution reconstruction techniques choose the Algebraic Iterative Reconstruction (ART) algorithm from image reconstruction algorithms. However, the ART algorithm has a large computational load and long reconstruction time, resulting in limited application areas. Therefore, this invention, based on the adopted TDLAS joint detection system, enhances the advantages of the TDLAS in-situ detection system by improving the traditional ART algorithm, incorporating in-situ correction, and proposing a joint correction algebraic iterative algorithm to accelerate reconstruction speed and improve reconstruction accuracy, such as... Figure 5 As shown, the specific steps include the following:

[0074] Step (1) Reconstruct the two-dimensional distribution map of gas concentration using an algebraic iterative algorithm:

[0075] To reconstruct a two-dimensional gas concentration distribution map using an algebraic iterative algorithm, the image first needs to be discretized. This involves discretizing the entire two-dimensional gas concentration distribution image f(x,y) into N grids, where gas parameters such as temperature, pressure, and concentration are considered uniform. The gas concentration within each grid is obtained by inverting the absorption coefficient, and the grid value f... i (1≤i≤N) represents the gas absorption coefficient α in the i-th grid. i According to Beer-Lambert's law, when a laser beam j passes through a target region, the integrated absorbance of the absorption line with center frequency v can be discretized as:

[0076]

[0077] Among them, A j Let S be the integrated absorbance of the j-th ray; S(T) is the linear intensity at temperature T.

[0078] The above equation is also known as the gas absorption equation, where M is the total number of laser beams; [PXS(T)] i L is the product of pressure, concentration, and linear intensity at grid point i; i,j is the projection coefficient, which is numerically equal to the length of the laser beam j intercepted by the grid i.

[0079] In actual testing, detection lasers need to be emitted from different directions and angles to scan and reconstruct the area. Each laser beam has a gas absorption equation as described in formula (1). These gas absorption equations are combined and expressed as a set of linear equations as shown in formula (2):

[0080]

[0081] At this point, the reconstructed absorption coefficient can be obtained by solving the high-dimensional linear equations. However, due to the instability of actual measurement scenarios, the effective detection data may be more or less than the dimension of the linear equations, leading to no solution or infinite solutions for the linear equations. Therefore, the reconstructed absorption coefficient cannot be obtained directly using analytical methods; iterative algorithms are needed to approximate the true absorption coefficient of each grid. According to the ART algorithm, the gas absorption coefficient vector is generally set as α = (α1, α2, ..., α...). N The initial value vector of ) is the zero vector. Substitute it into equation (2) for iterative calculation, and then use the gas absorption coefficient vector α. (j-1) Substituting into the j-th equation of equation (2), we obtain α. (j) The iterative formula is:

[0082]

[0083] Where λ (0 < λ < 2) is the relaxation factor, which is related to the convergence speed and reconstruction quality; n (1 ≤ n ≤ N) represents the index of the N-dimensional image vector. Following the above iteration, α is obtained from the Mth equation in equation (2). (M) If the result is positive, then a complete iteration is considered to have been completed. In each iteration, it is determined whether the change in the gas absorption coefficient vector is greater than the threshold (10). -6 And whether the number of iterations is less than 10,000. If equation (3) converges and the convergence speed is fast, the iteration will end early before 10,000 iterations because the change in the absorption vector is less than the threshold. If equation (3) converges but the convergence speed is slow or equation (3) does not converge, the iteration will end because it exceeds 10,000 iterations. Then, the concentration distribution can be solved by looking up the linear intensity at room temperature in the Hisran database.

[0084] Step (2) Design an adaptive correction algebraic iteration algorithm:

[0085] As the number of laser beams increases, the dimension of the linear equation system also increases, leading to increased computational complexity and reconstruction time. To address this, the Modified Adaptive Algebraic Reconstruction Technique (MAART) is employed. This algorithm adaptively adjusts the relaxation factor, improving reconstruction accuracy and reducing reconstruction time. The gas absorption coefficient α within each grid cell is also addressed. i The value can be obtained by solving the following iterative formula.

[0086]

[0087]

[0088] j k =1+[k mod(M)] (4c)

[0089] Where k represents the number of iterations in the MAART iteration process; β is related to the adaptive adjustment step size, typically taking a value of 0.25; Let be the adaptive relaxation factor, representing the relaxation factor of grid i in the k-th iteration; mod() is the remainder function. As shown in Equation 4(b), the greater the contribution of grid i's absorption to the laser integrated absorbance, the larger the adaptive relaxation factor of grid i. Linking the magnitude of the adaptive relaxation factor to the contribution of grid i's absorption to the laser integrated absorbance not only achieves fast convergence but also high-quality reconstruction.

[0090] Step (3) uses nonnegativity constraints and smoothness criteria to constrain the reconstruction process:

[0091] This invention employs non-negativity constraints and a smoothing criterion to constrain the reconstruction process, making the reconstruction results closer to the actual gas concentration distribution and reducing fluctuations in the reconstruction results. A non-negativity constraint is added to the gas absorption coefficient, and the non-negativity constraint function Φ() is shown in equation (5):

[0092]

[0093] At the same time, a smoothing criterion is used to suppress abrupt changes in the reconstruction results. In each iteration, the absorption coefficient of grid i is smoothed using itself and 8 surrounding grids, a total of 9 points. The smoothing criterion is shown in equation (6):

[0094]

[0095] Where δ is the smoothing factor, typically taking the value of 0.01, and m and n in the above formula are the row and column indices of the absorption coefficient in the reconstructed region grid, respectively.

[0096] To eliminate the influence of pressure P and linear strength S(T) on the iteration results, a new weighting factor L' is defined. i,j :

[0097] L' i,j =P i ·L i,j ·S i (T i (7)

[0098] Among them, S i (T i ) represents the line strength of the i-th grid.

[0099] The integral absorbance, or projection value A, at this point can be expressed as:

[0100] A = X·L' (8)

[0101] Where L′ is the new weighting factor;

[0102] For each absorption line, if there are I projected rays, there will be I absorption equations. The original linear equations about the absorption coefficient are transformed into a linear equation system about the concentration X.

[0103] Step (4) Solve the gas concentration field using the combined modified algebraic iterative algorithm:

[0104] To further improve reconstruction accuracy, this invention proposes a joint correction algebraic iterative algorithm to solve the gas concentration field. In actual measurements, based on the mobile TDLAS joint detection system, not only are remotely measured gas concentration values ​​obtained, but also in-situ gas concentration values ​​are obtained. For the gas absorption coefficient at the grid where the in-situ measurement is located, an in-situ correction is added in each iteration. The in-situ correction formula is as follows, where the correction factor χ = 0.01:

[0105]

[0106] In the above formula, α inSitu,i Let be the in-situ gas concentration measurement value at grid i. The detection accuracy of in-situ gas concentration measurement reaches the ppm level, or even the ppb level, which is much higher than the detection accuracy of telemetry gas concentration measurement. It is closer to the actual gas concentration value and the true value. Therefore, adding in-situ constraints will greatly improve the accuracy of concentration reconstruction and reduce the reconstruction time, because the number of unknowns in the equation is reduced.

[0107] The following is a comparison of the gas two-dimensional concentration distribution reconstruction results based on TDLAS joint detection and TDLAS telemetry:

[0108] This invention studies the reconstruction results based on different detection methods under the three Gaussian peak model. A 20m × 20m enclosed room is divided into a 20 × 20 grid, and the detection route is as follows: Figure 2 As shown, the detection interval of the TDLAS telemetry system and the TDLAS in-situ detection system is reduced to 0.125s, the number of rays in the TDLAS telemetry system increases to 876, and the number of detection points in the TDLAS in-situ detection system also increases to 876.

[0109] Figure 6a , Figure 6b , Figure 6c , Figure 6d , Figure 6e , Figure 6f , Figure 6g , Figure 6h The images show reconstruction results based on TDLAS telemetry and reconstruction results based on TDLAS joint detection; among them, Figure 6a A 3D image of the reconstruction result from the ART algorithm. Figure 6b A top view of the reconstruction results from the ART algorithm. Figure 6c A stereoscopic image of the reconstruction result from the MAART algorithm. Figure 6d A top view of the reconstruction results from the MAART algorithm. Figure 6e To reconstruct the image using the MAART algorithm with non-negativity constraints and smoothing criteria, Figure 6f The reconstruction result of the MAART algorithm with non-negativity constraints and smoothing criteria added; Figure 6g To reconstruct the 3D image using the joint modified algebraic iterative algorithm, Figure 6hA top view of the reconstruction results of the joint modified algebraic iterative algorithm.

[0110] According to the reconstruction results, the ART algorithm's reconstructed image differs the most from the three-Gauss peak model, and the reconstructed image is not smooth enough, especially in the region where the original concentration change is gentle, producing a lot of oscillation waves; the MAART algorithm improves the reconstruction quality to a certain extent and has a certain suppression effect on oscillation waves; after adding non-negativity constraints and smoothing criteria, the reconstructed image is smoother, but the radius of the oscillation waves increases, reducing the reconstruction quality; the joint modified algebraic iterative algorithm significantly improves the reconstruction quality, greatly reduces oscillation waves, and the reconstructed image is smooth, almost identical to the three-Gauss peak model. In order to quantitatively analyze the reconstruction accuracy of each reconstruction algorithm, this invention calculates the relative root mean square error (RRMSE) of each reconstruction algorithm according to formula (10), as shown in Table 1.

[0111]

[0112] Among them, C rec,i and C ori,i These represent the reconstructed concentration value and the numerical simulation value of the i-th grid, respectively; is the mean of the numerical simulation values ​​for all grids; n is the total number of grids.

[0113] Table 1

[0114]

[0115] According to the reconstruction error table, the ART algorithm has the largest reconstruction error, with an RRMSE of 5.028%. Adaptively adjusting the relaxation factor reduces the reconstruction error, with the MAART algorithm having an RRMSE of 2.537%. Adding non-negativity constraints and smoothing criteria can suppress non-smoothness, but it also changes some meshes with better reconstruction results, thus increasing the reconstruction error to 3.586%. Combining in-situ detection results greatly reduces the reconstruction error, which is about 0.582 times that of the ART algorithm, and the combined modified algebraic iterative algorithm has an RRMSE of 2.926%.

[0116] Table 2 shows the reconstruction time of each reconstruction algorithm. The ART algorithm has the longest reconstruction time of 127.430s. The reconstruction time of the MAART algorithm is reduced to 96.176s after adaptively adjusting the size of the relaxation factor. The reconstruction time is further reduced to 50.636s after adding non-negativity constraints and smoothing criteria. The reconstruction time is reduced to 39.267s after combining the in-situ detection results, which is 0.308 times that of the ART algorithm.

[0117] Table 2

[0118]

[0119] In summary, when the gas concentration distribution model is complex, the joint modified algebraic iterative algorithm has the smallest reconstruction error, the MAART algorithm has a smaller reconstruction error than the MAART algorithm with added non-negativity constraints and smoothing criteria, and the ART algorithm has the largest reconstruction error. The joint modified algebraic iterative algorithm has the fastest reconstruction speed, the MAART algorithm with added non-negativity constraints and smoothing criteria is faster than the MAART algorithm, and the ART algorithm is the slowest. Therefore, under different gas concentration distribution models, adaptively adjusting the relaxation factor and combining in-situ detection results can reduce reconstruction error and reconstruction time, and better reflect the gas concentration distribution trend. Adding non-negativity constraints and smoothing criteria introduces some error, but can suppress image roughness and reduce reconstruction time. Among these, the joint modified algebraic iterative algorithm produces the best reconstruction results; therefore, the joint detection method yields better reconstruction results than the telemetry method alone.

[0120] Those skilled in the art will readily understand that 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 scope of protection of the present invention.

Claims

1. A TDLAS combined detection method, characterized in that, Includes the following steps: Step 1: Acquire TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment. Measure indoor gas concentration data using the TDLAS telemetry system and TDLAS in-situ detection system along the detection route. Step 2: Reconstruct the two-dimensional concentration distribution map of indoor gases. Using TDLAS telemetry data and TDLAS in-situ detection data of indoor gas environment, the two-dimensional concentration distribution map of indoor gases is reconstructed according to the joint modified algebraic iterative algorithm. Reconstructing the two-dimensional concentration distribution map of indoor gases includes: The two-dimensional concentration distribution map of indoor gases is reconstructed using a joint modified algebraic iterative algorithm, which includes the following steps: Step (1) Reconstruct the two-dimensional distribution map of gas concentration using an algebraic iterative algorithm; Step (2) employs an adaptive correction algebraic iterative algorithm to improve reconstruction accuracy and speed; Step (3) uses non-negativity constraints and smoothness criteria to constrain the reconstruction process; Step (4) Design a joint modified algebraic iterative algorithm to further improve reconstruction accuracy and reconstruction speed, including: For the gas absorption coefficient at the grid where the in-situ measurement is located, an in-situ correction is added in each iteration. The in-situ correction formula is as follows, where the correction factor is... : (9) In the above formula For grid In-situ gas concentration measurement at the location.

2. The TDLAS combined detection method according to claim 1, characterized in that, In step 1, acquiring TDLAS telemetry data of the indoor gas environment includes: The TDLAS telemetry system is designed to detect gas concentrations along the detection route at a certain frequency.

3. The TDLAS combined detection method according to claim 2, characterized in that, In step 1, obtaining TDLAS in-situ detection data of the indoor gas environment includes: While the TDLAS telemetry system detects gas concentration at a certain frequency, the TDLAS in-situ detection system detects gas concentration at the same frequency.

4. The TDLAS combined detection method according to claim 1, characterized in that, In step (1), the image is first discretized, that is, the entire two-dimensional distribution image of gas concentration is obtained. Discretize into A grid is defined, and the gas parameters within the grid are considered uniform; the gas concentration within the grid is obtained by inversion of the gas absorption coefficient, and the grid value f is given. i Indicates the first Gas absorption coefficient within each grid , According to Beer-Lambert's law, when a certain laser beam... Passing through the target area, the center frequency is The integral absorbance of the absorption line is discrete as follows: (1) in, For the first The integral absorbance of a single ray; For temperature is Line strength at that time; The above equation is the gas absorption equation, where, This represents the total number of laser beams. For grid The product of pressure, concentration, and linear intensity; The projection coefficient is numerically equal to the laser beam. Grid The length of the cut-off section; In the actual detection process, detection lasers are emitted from different directions and angles to scan and reconstruct the area. Each laser beam has a gas absorption equation as described in equation (1). The gas absorption equations of each laser beam are combined and expressed as a set of linear equations as shown in equation (2): (2) The reconstructed value of the gas absorption coefficient is obtained by solving the linear equation system, and a gas absorption coefficient vector is set. The initial value vector is the zero vector. Substitute it into equation (2) for iterative calculation, and then calculate the gas absorption coefficient vector. Substituting into equation (2) When the equation is 1, we get The iterative formula is: (3) in, It is a relaxation factor. It is related to the convergence speed and reconstruction quality; n represents Index of a 3D image vector According to the above iteration, from equation (2) the... The equations yield Then, a complete iteration is completed. In each iteration, it is determined whether the change in the gas absorption coefficient vector is greater than the threshold and whether the number of iterations is less than 10000. Then, the gas concentration distribution is solved by finding the linear intensity at room temperature according to the Hisran database.

5. The TDLAS combined detection method according to claim 4, characterized in that, Step (2) includes: An adaptive correction algebraic iterative algorithm is used to determine the gas absorption coefficient within each grid cell. The value is obtained by solving the following iterative formula: (4a) (4b) (4c) in, This represents the number of iterations in the MAART iteration process; It is related to the adaptive adjustment step size, and takes a value of 0.25; Let be the adaptive relaxation factor, representing the th In the next iteration, the grid relaxation factor, For finding the remainder function.

6. The TDLAS combined detection method according to claim 5, characterized in that, Step (3) includes: To apply a nonnegativity constraint to the gas absorption coefficient, the nonnegativity constraint function is... As shown in equation (5): (5) Simultaneously, a smoothing criterion is used to suppress abrupt changes in the reconstruction results. In each iteration, the mesh... The gas absorption coefficient is smoothed using nine points, including the coefficient itself and eight surrounding grids. The smoothing criterion is shown in equation (6): (6) in, As the smoothing factor, in the above formula and These are the row and column indices of the absorption coefficient in the reconstructed region grid; To eliminate pressure and line strength To influence the iteration results, define a new weighting factor. : (7) in, For the first Line strength of each grid; The integral absorbance at this point is also the projected value. Represented as: (8) in, As the new weighting factor; For each absorption line, if there is A projected ray of light will have The absorption equations, a system of linear equations concerning the gas absorption coefficient, are transformed into equations concerning concentration. A system of linear equations.

Images