Boiler furnace three-dimensional combustion temperature field reconstruction method and system

Abstract

The invention discloses a boiler furnace three-dimensional combustion temperature field reconstruction method and system, and the method comprises the steps: S1, multi-mode sensing network deployment, S2, time-space synchronization data collection, S3, sound wave flying time extraction and acoustic modeling, S4, flame image processing and radiation temperature conversion, S5, three-dimensional temperature initial field generation, and S6, multi-mode joint objective function construction. S7, carrying out regularization constraint iteration solution; and S8, carrying out three-dimensional temperature field visualization output. According to the method, high penetrability and sensitivity of sound waves to medium temperature and rich radiation information contained in flame images are fully utilized, and the precision and spatial resolution of temperature field reconstruction are remarkably improved by establishing a joint objective function and performing collaborative optimization. Meanwhile, a regularization method and an adaptive filtering technology are introduced, interference caused by measurement noise and uncertain problems is effectively suppressed, and the stability and robustness of the reconstruction process are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of boilers, in particular to a boiler furnace three-dimensional combustion temperature field reconstruction method and system. BACKGROUND

[0002] In the field of energy and power engineering, the three-dimensional combustion temperature field in the boiler furnace is a key parameter for evaluating combustion efficiency, pollutant emission and equipment safety.

[0003] Traditional temperature measurement methods, such as thermocouples and infrared temperature measurement, can usually only provide single-point or two-dimensional temperature information, making it difficult to fully reflect the complex and variable three-dimensional combustion conditions inside the furnace. Although non-contact methods such as acoustic temperature measurement and radiation temperature measurement have improved, they often face problems such as low reconstruction accuracy, poor anti-interference ability, and poor real-time performance when reconstructing three-dimensional temperature fields, especially in harsh boiler combustion environments, where noise and measurement errors can severely affect the accuracy and reliability of temperature field reconstruction.

[0004] Therefore, there is an urgent need for a three-dimensional combustion temperature field reconstruction technology that can achieve high precision, high robustness and real-time visualization. SUMMARY

[0005] To this end, the present application provides a boiler furnace three-dimensional combustion temperature field reconstruction method and system to solve the problems in the prior art.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0007] A boiler furnace three-dimensional combustion temperature field reconstruction method and system, comprising the following steps:

[0008] S1, multi-modal sensor network deployment, arranging a plurality of acoustic transducer arrays and a plurality of color CCD industrial cameras in the boiler furnace to construct a multi-modal data acquisition system; wherein the acoustic transducer array is used to emit and receive acoustic signals, and the plurality of color CCD industrial cameras are used to synchronously capture images of the combustion flame in the furnace from different angles;

[0009] S2, time and space synchronous data acquisition, controlling the acoustic transducer array to work according to a predetermined time sequence, sequentially exciting acoustic signals and recording the time of flight, and simultaneously triggering all color CCD industrial cameras to synchronously capture images to obtain an acoustic time of flight dataset and a furnace flame image dataset at the same time;

[0010] S3, acoustic time of flight extraction and acoustic modeling, preprocessing the collected original acoustic signals; based on the acoustic time of flight, combining a pre-established furnace space acoustic propagation path model to construct a linear equation set for acoustic tomography;

[0011] S4, flame image processing and radiation temperature conversion, pre-processing the collected furnace flame image; based on the three primary color temperature measurement method, extracting the RGB value representing the temperature information from the corrected flame image, and converting the RGB value into two-dimensional temperature distribution information according to the color and temperature mapping relationship obtained by pre-calibration through the blackbody furnace;

[0012] S5, three-dimensional temperature initial field generation, taking the two-dimensional temperature distribution information obtained in step S4 as an initial iterative field, converting it from a two-dimensional image space to a three-dimensional furnace space through a coordinate mapping relationship, and generating an initial estimate of a three-dimensional temperature distribution;

[0013] S6, multi-modal joint objective function construction, constructing a joint reconstruction objective function, which includes an acoustic reconstruction term and an image reconstruction term; the acoustic reconstruction term is the two-norm of the difference between the theoretical flight time calculated based on the acoustic tomography equation group obtained in step S3 and the actual measured flight time; the image reconstruction term is the two-norm of the difference between the theoretical radiation value calculated based on the current iterative temperature field through the radiation transfer model and the actual image radiation value extracted in step S4;

[0014] S7, regularized constraint iterative solution, using a Landweber iteration algorithm with regularization constraint to solve the joint reconstruction objective function, in the iteration process, Tikhonov regularization is introduced to suppress the ill-posedness of the solution, and the conjugate gradient method is used to accelerate the solution process, until the preset convergence condition is met, and finally the three-dimensional combustion temperature field distribution data of the boiler furnace is output;

[0015] S8, three-dimensional temperature field visualization output, inputting the reconstructed three-dimensional combustion temperature field distribution data into a three-dimensional visualization module, and generating a three-dimensional temperature field cloud chart of the furnace interior on a display device through volume rendering technology.

[0016] Preferably, in step S1, the arrangement of the acoustic transducer array is: uniformly arranged in a ring shape at four different height layers of the furnace, 8 acoustic transducers are arranged at each layer, and the transducers between adjacent layers are staggered by 45 degrees in the vertical direction.

[0017] Preferably, the number of the plurality of color CCD industrial cameras is 4, which are arranged at the fire observation holes of the four walls of the furnace, and the optical axes of all the cameras form an elevation angle of 10 to 30 degrees with the horizontal plane.

[0018] Preferably, the pre-established acoustic propagation path model of the furnace space is obtained in the following way: a three-dimensional geometric model of the boiler furnace is established, and the three-dimensional geometric model is discretized into several voxel units in three-dimensional space; for each pair of sound wave transmitting and receiving transducers, the ray tracing algorithm is used to simulate the path of the sound wave from the transmitting point to the receiving point, and the effective length of each voxel unit traversed by the path is calculated, thereby forming the weight coefficient vector of the path. The weight coefficient vectors of all sound wave paths together constitute the projection matrix of acoustic tomography.

[0019] Preferably, in step S6, the joint reconstruction objective function takes the following form:

[0020]

[0021] in, Let be the three-dimensional temperature field vector to be determined. For acoustic projection matrix, The measured sound wave transit time vector. For temperature field The theoretical image radiance vector calculated using the radiative transfer model. This is the vector of radiometric values ​​from the actually acquired and processed image. The weight matrix is ​​constructed based on the camera position and the direction of the gaze. These are the weighting coefficients for the image items. The weight coefficients for the regularization term are... This is a Tikhonov regularization term.

[0022] Preferably, in step S7, the iterative format of the Landweber iterative algorithm with regularization constraints is:

[0023]

[0024] in, For the first Temperature field of the next iteration The iteration step size, Acoustic projection matrix transpose, For radiative transfer model exist Jacobian matrix at the location, This is the Tikhonov regularization matrix.

[0025] To achieve the above objectives, the present invention also provides the following technical solution:

[0026] A three-dimensional combustion temperature field reconstruction system for boiler furnaces includes:

[0027] a data acquisition module comprising an acoustic transducer array and a color CCD industrial camera for synchronously acquiring acoustic signals and flame image signals in the furnace;

[0028] a signal preprocessing module, electrically connected with the data acquisition module, for carrying out wavelet denoising on the acquired original acoustic signals to extract the flyover time and carrying out denoising and geometric correction on the acquired original flame images;

[0029] an acoustic reconstruction model construction module, connected with the signal preprocessing module, for constructing an acoustic tomography equation set based on the acoustic flyover time and a pre-established acoustic propagation path model of the furnace space;

[0030] an image temperature conversion module, connected with the signal preprocessing module, for converting the preprocessed flame images into two-dimensional temperature distribution information based on a three-primary-color temperature measurement method and a calibrated color-temperature mapping relationship;

[0031] an initial field generation module, connected with the image temperature conversion module, for converting the two-dimensional temperature distribution information into an initial temperature field estimation value of the three-dimensional furnace space through coordinate mapping;

[0032] a multi-modal fusion reconstruction module, connected with the acoustic reconstruction model construction module and the initial field generation module respectively, which internally has a joint reconstruction objective function and a solving algorithm, for fusing acoustic and image information through iterative calculation to reconstruct a high-precision three-dimensional temperature field;

[0033] a three-dimensional visualization module, connected with the multi-modal fusion reconstruction module, for generating a three-dimensional temperature field cloud image through volume rendering technology and displaying the three-dimensional temperature field data obtained through reconstruction.

[0034] The present application has the following advantages: the present application makes full use of the high penetration and sensitivity of acoustic waves to medium temperature and the rich radiation information contained in flame images, significantly improves the accuracy and spatial resolution of temperature field reconstruction through the establishment of a joint objective function and collaborative optimization. At the same time, the introduction of regularization method and adaptive filtering technology effectively suppresses the interference brought by measurement noise and ill-posed problem, enhances the stability and robustness of the reconstruction process.

[0035] The complete system constructed by the present application realizes the full-process automation from data acquisition, processing to three-dimensional visualization, can provide accurate data support and decision basis for the optimized combustion of the boiler, the reduction of nitrogen oxide emissions and the prevention of coking and corrosion, and has great engineering application value. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more intuitively illustrate the prior art and the present application, exemplary drawings are given below. It should be understood that the specific shapes, structures shown in the drawings should not be regarded as limiting conditions in the implementation of the present application; for example, based on the technical concepts disclosed in the present application and the exemplary drawings, those skilled in the art can easily make routine adjustments or further optimizations to the increase / decrease / assignment of certain units (components), specific shapes, positional relationships, connection methods, size ratio relationships, etc.

[0037] Figure 1 An implementation flowchart of a boiler furnace three-dimensional combustion temperature field reconstruction method provided for an embodiment of the present application.

[0038] Figure 2 A module diagram of a boiler furnace three-dimensional combustion temperature field reconstruction system of the present application. DETAILED DESCRIPTION

[0039] The embodiments of the present application are described below by specific specific embodiments, and those skilled in the art can easily understand other advantages and effects of the present application from the disclosed content. Obviously, the described embodiments are part of the embodiments of the present application, not all. It should be understood that these embodiments are only for further illustration of the present application, and cannot be understood as limiting the scope of protection of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0040] Please refer to Figure 1 A boiler furnace three-dimensional combustion temperature field reconstruction method and system, characterized in that it comprises the following steps:

[0041] S1, multi-modal sensor network deployment, a plurality of acoustic transducer arrays and a plurality of color CCD industrial cameras are arranged on the circumferential wall of the boiler furnace to construct a multi-modal data acquisition system; wherein the acoustic transducer array is used to emit and receive acoustic signals, and the plurality of color CCD industrial cameras synchronously acquire images of the combustion flame in the furnace from different angles. In order to ensure that the acoustic transducer and the color CCD industrial camera can work normally, a protective structure and a cooling system can be provided during implementation, which is not limited here;

[0042] S2, time-space synchronous data acquisition, control the acoustic transducer array to work according to the predetermined time sequence, sequentially excite acoustic signals and record the flying time, and at the same time trigger all color CCD industrial cameras to synchronously capture images, so as to obtain the acoustic flying time data set and the furnace flame image data set at the same time;

[0043] S3, acoustic flight time extraction and acoustic modeling, preprocessing the collected original acoustic signal, including wavelet denoising and signal enhancement, to extract accurate acoustic flight time; based on the acoustic flight time, combining the pre-established acoustic propagation path model of the furnace space, constructing the linear equation set of acoustic tomography, which describes the relationship between the acoustic flight time and the temperature integral along the propagation path;

[0044] S4, flame image processing and radiation temperature conversion, preprocessing the collected furnace flame image, including image denoising, background removal and geometric correction; based on the three primary color temperature measurement method, extracting the RGB value representing temperature information from the corrected flame image, and converting the RGB value to two-dimensional temperature distribution information according to the color-temperature mapping relationship obtained by pre-calibration of the blackbody furnace;

[0045] S5, three-dimensional temperature initial field generation, taking the two-dimensional temperature distribution information obtained in step S4 as the initial iterative field, converting it from the two-dimensional image space to the three-dimensional furnace space through the coordinate mapping relationship, and generating an initial estimate of the three-dimensional temperature distribution;

[0046] S6, multi-modal joint objective function construction, constructing a joint reconstruction objective function, which includes an acoustic reconstruction term and an image reconstruction term; the acoustic reconstruction term is the two-norm of the difference between the theoretical flight time calculated based on the acoustic tomography equation set obtained in step S3 and the actual measured flight time; the image reconstruction term is the two-norm of the difference between the theoretical radiation value calculated based on the current iterative temperature field through the radiation transfer model and the actual image radiation value extracted in step S4;

[0047] S7, regularized constraint iterative solution, using the Landweber iteration algorithm with regularization constraint to solve the joint reconstruction objective function, in the iteration process, Tikhonov regularization is introduced to suppress the ill-posedness of the solution, and the conjugate gradient method is used to accelerate the solution process, until the preset convergence condition is met, and finally output the high-precision three-dimensional combustion temperature field distribution data of the boiler furnace;

[0048] S8, three-dimensional temperature field visualization output, inputting the reconstructed three-dimensional combustion temperature field distribution data into the three-dimensional visualization module, generating a three-dimensional temperature field cloud chart of the furnace interior on the display device through volume rendering technology, realizing real-time, dynamic and three-dimensional visualization of the combustion temperature field.

[0049] In the implementation of the present scheme, firstly, the acoustic and optical physical signals at the same time in the furnace are synchronously collected by the acoustic transducer and the CCD camera. The acoustic flight time carries the temperature integral information along its propagation path, while the flame image directly reflects the two-dimensional radiation intensity distribution of the combustion products. The present scheme converts the image information into a two-dimensional temperature field through the three primary color method, and takes it as the initial field, which is mapped from the two-dimensional image space to the three-dimensional furnace space, providing a physically reasonable starting point for subsequent reconstruction, avoiding the blindness of traditional methods starting from zero iteration. The most critical step is to construct a target function that combines the acoustic reconstruction term and the image reconstruction term, which unifies the integral temperature measurement value along the acoustic path and the radiation intensity measurement value of the image pixel points in a mathematical framework. Then the regularization constrained iterative algorithm (such as Landweber algorithm) is introduced to solve the target function, which is essentially to make the calculated three-dimensional temperature field satisfy the physical constraints of acoustic measurement and image measurement at the same time. Through the complementation and mutual correction of the two kinds of information, a more accurate and stable three-dimensional combustion temperature field is finally reconstructed than single method.

[0050] The acoustic transducers are arranged in a ring shape and uniformly on four different height layers in the furnace, with 8 acoustic transducers on each layer. The transducers on adjacent layers are staggered by 45 degrees in the vertical direction to optimize the coverage of the acoustic path in the three-dimensional space of the furnace.

[0051] By arranging the acoustic transducers in a ring shape and staggering the angles at different height layers, the cross-coverage density and angle of the acoustic path in the three-dimensional space can be significantly increased. This "three-dimensional grid" type path layout greatly improves the mathematical conditions of the reconstruction problem (i.e. the properties of the projection matrix), thereby effectively improving the spatial resolution of the temperature field in the vertical and horizontal directions. For example, a 4-layer × 8 staggered arrangement can more accurately detect the temperature difference between the combustion zone at the bottom of the furnace and the burnout zone at the top, compared to a single-layer 16 planar arrangement.

[0052] The number of the plurality of color CCD industrial cameras is 4, which are arranged at the fire observation holes of the four walls of the furnace. The arrangement of the four cameras ensures that the entire cross-section of the furnace can be covered from multiple angles, avoiding visual dead angles, and the optical axes of all cameras are at an elevation angle of 10 to 30 degrees with the horizontal plane to ensure that the flame image of the central area of the furnace can be captured.

[0053] The pre-established furnace space acoustic propagation path model is obtained by: establishing a precise three-dimensional geometric model of the boiler furnace, discretizing the model in three-dimensional space into a plurality of voxel units; for each pair of sound wave emission and receiving transducer, using a ray tracing algorithm to simulate the path of the sound wave from the emission point to the receiving point, and calculating the effective length of each voxel unit through which the path passes, thereby forming a weight coefficient vector of the path, and the weight coefficient vectors of all sound wave paths collectively constitute a projection matrix for acoustic tomography.

[0054] Discretizing the continuous physical space into a voxel grid and accurately calculating the specific voxel through which each sound wave path passes and its length using a ray tracing technique converts the continuous integral relationship into a linear equation set (L x T = t), where each element of the projection matrix L represents the spatial relationship between the path and the voxel, which is the cornerstone of converting physical problems into calculable mathematical models.

[0055] The specific form of the joint reconstruction objective function is:

[0056]

[0057] wherein, is a three-dimensional temperature field vector to be solved, is an acoustic projection matrix, is a measured sound wave flight time vector, is a theoretical image radiation value vector calculated based on the temperature field by a radiation transfer model, is an image radiation value vector actually collected and processed, is a weight matrix constructed based on the camera position and the line of sight direction, is a weight coefficient of the image term, is a weight coefficient of the regularization term, is a Tikhonov regularization term.

[0058] The weight coefficients and balance the contribution proportion of the acoustic data, the image data, and the prior constraint (smoothness) of the solution in the final solution. For example, when the image quality is poor due to dust interference, the value of a can be adjusted to be low, and more reliance is placed on acoustic data; conversely, the weight of the image data can be increased to achieve adaptive optimization. That is, when the quality of a certain type of data (such as the image due to the reduction of the signal-to-noise ratio caused by dust interference) decreases, the weight coefficient can be adjusted to make the reconstruction process rely more on another type of more reliable data, thereby significantly improving the robustness and adaptability of the entire system in a complex industrial environment, and ensuring the stability and reliability of the reconstruction result.

[0059] In step S7, the iteration format of the Landweber iteration algorithm with regularization constraint is as follows:

[0060]

[0061] wherein, is the temperature field of the i-th iteration, is the iteration step size, is the transpose of the acoustic projection matrix is the Jacobian matrix at is the Tikhonov regularization matrix, which forces the solution to satisfy a certain smoothness, prevents the generation of non-physical sharp oscillations due to noise, avoids the appearance of non-physical sharp fluctuations, and improves the reconstruction quality. As shown in , a three-dimensional combustion temperature field reconstruction system for a boiler furnace comprises: A data acquisition module comprising an array of acoustic transducers and a color CCD industrial camera is used to synchronously acquire acoustic signals and flame image signals in the furnace;

[0062] Figure 2 A signal preprocessing module is electrically connected to the data acquisition module and is used to perform wavelet denoising on the acquired original acoustic signals to extract accurate flyover times and to perform denoising and geometric correction on the acquired original flame images;

[0063] An acoustic reconstruction model construction module is connected to the signal preprocessing module and is used to construct an acoustic tomography equation set based on the acoustic flyover times and a pre-established acoustic propagation path model of the furnace space;

[0064] An image temperature conversion module is connected to the signal preprocessing module and is used to convert the preprocessed flame images into two-dimensional temperature distribution information based on a three-primary-color temperature measurement method and a calibrated color-temperature mapping relationship;

[0065] An initial field generation module is connected to the image temperature conversion module and is used to convert the two-dimensional temperature distribution information into an initial temperature field estimate value of the three-dimensional furnace space through coordinate mapping;

[0066] A multi-modal fusion reconstruction module is connected to the acoustic reconstruction model construction module and the initial field generation module, respectively, and has a built-in joint reconstruction objective function and a solving algorithm, and is used to fuse acoustic and image information through iterative calculation to reconstruct a high-precision three-dimensional temperature field;

[0067]

[0068]

[0069] ​​​​A three-dimensional visualization module, connected with the multi-modal fusion reconstruction module, is configured to generate and display a three-dimensional temperature field cloud chart through volume rendering technology.

[0070] In the implementation of the system, the data acquisition module synchronously triggers the acoustic transducer array and the industrial camera under the instruction of the central synchronization controller to capture the original acoustic signal and flame image. Subsequently, the signal preprocessing module processes the two signals in parallel: one is to perform wavelet denoising on the acoustic signal to extract the accurate flyover time, and the other is to perform denoising and geometric correction on the flame image. The processed data are respectively sent to the acoustic reconstruction model construction module and the image temperature conversion module, the former generates an acoustic tomography equation, and the latter outputs a two-dimensional temperature map. The initial field generation module receives the two-dimensional temperature map and converts it into a three-dimensional initial temperature field. The multi-modal fusion reconstruction module, as the key of the system, takes the initial temperature field and the acoustic equation as inputs, runs a complex fusion reconstruction algorithm, and outputs the final high-precision three-dimensional temperature field data through iterative calculation. Finally, the three-dimensional visualization module receives these data, generates and displays an intuitive three-dimensional temperature field cloud chart in real time through GPU-accelerated volume rendering technology, and completes the whole-chain reconstruction from physical signals to visual information.

[0071] The above merely describes the preferred embodiments of the present application and is not used to limit the present application. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method and system for reconstructing a three-dimensional combustion temperature field in a boiler furnace, characterized by, The method comprises the following steps: S1, multi-modal sensor network deployment, arranging a plurality of acoustic transducer arrays and a plurality of color CCD industrial cameras in the boiler furnace to construct a multi-modal data acquisition system; wherein the acoustic transducer array is used for transmitting and receiving acoustic signals, and the plurality of color CCD industrial cameras synchronously collect images of the combustion flame in the furnace from different angles; S2, time-space synchronous data acquisition, controlling the acoustic transducer array to work according to a predetermined time sequence, sequentially exciting acoustic signals and recording the flying time, and at the same time triggering all color CCD industrial cameras to synchronously capture images, so as to obtain an acoustic flying time data set and a furnace flame image data set at the same time; S3, acoustic flying time extraction and acoustic modeling, preprocessing the collected original acoustic signals; based on the acoustic flying time, combining a pre-established acoustic propagation path model of the furnace space, a linear equation set of acoustic tomography is constructed; S4, flame image processing and radiation temperature conversion, preprocessing the collected furnace flame images; based on the three-primary-color temperature measurement method, the RGB values representing temperature information are extracted from the corrected flame images, and according to the color-temperature mapping relationship obtained by pre-calibration of a blackbody furnace, the RGB values are converted into two-dimensional temperature distribution information; S5, three-dimensional temperature initial field generation, taking the two-dimensional temperature distribution information obtained in step S4 as an initial iterative field, converting it from a two-dimensional image space to a three-dimensional furnace space through a coordinate mapping relationship, and generating an initial estimated value of a three-dimensional temperature distribution; S6, multi-modal joint objective function construction, a joint reconstruction objective function is constructed, which includes an acoustic reconstruction item and an image reconstruction item; the acoustic reconstruction item is the two-norm of the difference between the theoretical flying time calculated based on the acoustic tomography equation set obtained in step S3 and the actual measured flying time; the image reconstruction item is the two-norm of the difference between the theoretical radiation value calculated based on the current iterative temperature field through a radiation transmission model and the actual image radiation value extracted in step S4; S7, regularized constraint iterative solution, the joint reconstruction objective function is solved by using a Landweber iteration algorithm with regularization constraint, in the iteration process, Tikhonov regularization is introduced to suppress the ill-posedness of the solution, and the conjugate gradient method is used to accelerate the solution process, until the preset convergence condition is met, and finally the three-dimensional combustion temperature field distribution data of the boiler furnace is output; S8, three-dimensional temperature field visualization output, inputting the reconstructed three-dimensional combustion temperature field distribution data into a three-dimensional visualization module, and generating a three-dimensional temperature field cloud chart of the furnace interior on a display device through volume rendering technology.

2. The method according to claim 1, characterized in that, In step S1, the acoustic transducer array is arranged in a ring shape and uniformly arranged at four different height layers of the furnace, 8 acoustic transducers are arranged at each layer, and the transducers between adjacent layers are staggered by 45 degrees in the vertical direction.

3. The method according to claim 2, wherein The number of the plurality of color CCD industrial cameras is 4, which are arranged at the fire observation holes of the four walls of the furnace, and the optical axes of all the cameras form an elevation angle of 10 to 30 degrees with the horizontal plane.

4. The method according to claim 1, wherein, The pre-established furnace space acoustic propagation path model is obtained by: establishing a three-dimensional geometric model of the boiler furnace, discretizing the three-dimensional geometric model into a plurality of voxel units in a three-dimensional space; for each pair of sound wave emission and receiving transducer, using a ray tracing algorithm to simulate the path of the sound wave from the emission point to the receiving point, and calculating the effective length of each voxel unit through which the path passes, thereby forming a weight coefficient vector of the path, and the weight coefficient vectors of all sound wave paths collectively constitute a projection matrix of acoustic tomography.

5. The method according to claim 4, wherein In step S6, the specific form of the joint reconstruction objective function is: ； wherein, is the three-dimensional temperature field vector to be solved, is the acoustic projection matrix, is the measured acoustic travel time vector, is the temperature field based on the measured acoustic travel time vector, is the theoretical image radiance vector calculated by the radiative transfer model, is the actual acquired and processed image radiance vector, is the weight matrix constructed based on the camera positions and the line of sight directions, is the weight coefficient of the image term, is the weight coefficient of the regularization term, is the Tikhonov regularization term.

6. The method according to claim 5, wherein, In step S7, the iteration format of the Landweber iteration algorithm with regularization constraint is: ； in, For the first Temperature field of the next iteration The iteration step size, Acoustic projection matrix transpose, For radiative transfer model exist Jacobian matrix at the location, This is the Tikhonov regularization matrix.

7. A system for reconstructing a three-dimensional combustion temperature field of a boiler furnace, characterized by Comprising: A data acquisition module comprising an array of sound wave transducers and a color CCD industrial camera, for synchronously acquiring sound wave signals and flame image signals in the furnace; A signal preprocessing module electrically connected with the data acquisition module, for wavelet denoising the acquired original sound wave signals to extract the flyover time, and denoising and geometrically correcting the acquired original flame images; An acoustic reconstruction model construction module connected with the signal preprocessing module, for constructing an acoustic tomography equation set based on the sound wave flyover time and the pre-established furnace space acoustic propagation path model; An image temperature conversion module connected with the signal preprocessing module, for converting the preprocessed flame images into two-dimensional temperature distribution information based on a three-primary-color temperature measurement method and a calibrated color-temperature mapping relationship; An initial field generation module connected with the image temperature conversion module, for converting the two-dimensional temperature distribution information into an initial temperature field estimate of the three-dimensional furnace space through coordinate mapping; A multi-modal fusion reconstruction module connected with the acoustic reconstruction model construction module and the initial field generation module, which internally has a joint reconstruction objective function and a solving algorithm, for fusing acoustic and image information through iterative calculation to reconstruct a high-precision three-dimensional temperature field; A three-dimensional visualization module connected with the multi-modal fusion reconstruction module, for generating a three-dimensional temperature field cloud image of the reconstructed three-dimensional temperature field data through volume rendering technology and displaying the same.

Images