Flame combustion three-dimensional temperature and concentration detection method and device based on hyperspectral camera and polarization technology

Abstract

The invention discloses a flame combustion three-dimensional temperature concentration detection method and device based on a hyperspectral camera and a polarization technology, and belongs to the technical field of high-temperature flame combustion diagnosis and the technical field of optical measurement. The problem that high-precision collaborative measurement of temperature and gas component concentration cannot be realized by an existing spectrum-based single detection mode is solved. The method comprises the following steps: acquiring radiation spectrum information of flame and light intensity information in different polarization states from a plurality of visual angles and positions by using a hyperspectral camera and a polarization camera cooperatively, and forming a flame field three-dimensional data set by using the acquired information; based on the radiation transfer theory, the Planck law and the gas absorption spectrum are adopted to follow the Beer-Lambert law to invert the radiation data of the temperature field and the gas concentration field of the flame; establishing a three-dimensional scalar field of flame temperature and gas concentration by adopting a tomography method; and establishing a target function based on the information, and carrying out iterative optimization on the target function by adopting an optimization method based on least squares to obtain flame combustion three-dimensional temperature and different gas component concentration data.

Description

Technical Field

[0001] The present invention belongs to the technical fields of high-temperature flame combustion diagnosis and optical measurement technology. Background Art

[0002] Combustion is an efficient way to convert fossil energy into available heat energy. In order to improve the efficiency of combustion-based power plants, reduce the environmental impact of pollutant emissions, and optimize the structural design of power machinery, it is necessary to deeply understand the combustion process and achieve precise control. High-precision parameter field measurement technology is an important means for combustion research, providing key experimental data support for comprehensively revealing complex combustion mechanisms, improving combustor designs, and reducing pollutant emissions. The temperature field of the flame, the concentration distributions of products such as CO2, H2O, CO, etc., as well as the soot particle concentration and spectral radiation characteristics, directly reflect the basic characteristics of the combustion state. At the same time, these key physical quantities also characterize the chemical reaction rate inside the flame and the pollutant generation law. Therefore, deeply mastering the distributions of the three-dimensional temperature field, gas concentration field, and radiation characteristic field inside the flame is crucial for the optimal evaluation of combustion performance, pollutant control, and safety monitoring of the combustion process.

[0003] Currently, for the combustion diagnosis technology of high-temperature flames, non-contact measurement methods are mainly adopted, which are generally divided into two types: one is passive tomography detection based on radiation images, and the other is active tomography detection based on laser spectroscopy. The passive tomography detection method utilizes the information radiated by the flame itself, and these radiation signals contain important data inside the flame. By using detectors such as CCD cameras or external thermal imagers, these signals can be captured, and then the characteristics of the flame can be analyzed. The active tomography detection method obtains detection signals through the physical property changes generated when a laser passes through a high-temperature flame, usually by measuring the propagation of the laser in different directions to reconstruct the internal structure of the flame.

[0004] Due to the complex coupling of multiple physical fields inside the flame, the process of real-time reconstructing multiple physical fields faces difficulties such as "strong non-linearity", "high underdetermination", "severe crosstalk", and "large computational amount", which makes it extremely difficult to achieve three-dimensional real-time synchronous detection of multiple physical quantity fields of the flame. On the other hand, single subjective or passive tomography detection methods have many limitations, such as the limitations of detector layout and the known spectral radiation physical property parameters, resulting in their inability to accurately measure the three-dimensional spatial distributions of multiple parameters such as temperature, component concentration, and pressure of high-temperature flames. Summary of the Invention

[0005] The present invention is to solve the problem that the existing single detection method based on spectroscopy cannot achieve high-precision collaborative measurement of temperature and gas component concentration, and now provides a three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology.

[0006] The three-dimensional temperature and concentration detection method for flame combustion based on hyperspectral camera and polarization technology according to the present invention includes:

[0007] Step 1: Use a hyperspectral camera and a polarization camera to cooperatively collect the radiation spectrum information of the flame and the light intensity information under different polarization states from multiple perspectives and positions. Utilize the radiation spectrum information of the flame and the light intensity information under different polarization states to calculate the radiation data of multiple bands, obtain the three-dimensional spatial information of the flame, and form a three-dimensional data set of the flame field;

[0008] Step 2: Based on the radiation transfer theory, utilize the three-dimensional data set of the flame field, and adopt Planck's law and the Beer-Lambert law followed by the gas absorption spectrum to invert the radiation data of the temperature field and gas concentration field of the flame;

[0009] Step 3: Adopt a tomography method, utilize the spectral information of the flame, the light intensity information under different polarization states, and combine the radiation data of the temperature field and gas concentration field of the flame to establish a three-dimensional scalar field of the flame temperature and gas concentration;

[0010] Step 4: Establish an objective function by using the three-dimensional scalar field of the flame temperature and gas concentration and the spectral information of the flame or the light intensity information under different polarization states. Adopt an optimization method based on the least squares to iteratively optimize the three-dimensional scalar field of the flame temperature and gas concentration in the objective function, and determine in real time whether the objective function reaches the iteration stop condition. If so, obtain the numerical value of the three-dimensional scalar field of the corresponding flame temperature and gas concentration. Otherwise, return to Step 1 until the objective function reaches the iteration stop condition, and obtain the three-dimensional temperature of the flame combustion and the concentration data of different gas components.

[0011] Furthermore, in the present invention, in Step 2, the specific method for inverting the radiation data of the temperature field and gas concentration field of the flame is:

[0012] First, establish a transfer equation:

[0013]

[0014] where I λ is the radiation intensity at wavelength λ, κ λ is the gas absorption coefficient, σ λ is the scattering coefficient, J λ is the scattering source term, s is the path length along the light propagation direction, B λ is the blackbody radiation, and B λ (T) is obtained by Planck's law:

[0015]

[0016] where c1 is the first radiation constant, 3.7419×10-16 W·m 2 ; c2 is the second radiation constant, 1.4388×10 - 2 m·K;

[0017] Invert the temperature and gas concentration using the first and second terms of the transfer equation:

[0018]

[0019] Solve the temperature field T:

[0020]

[0021] where I λ1 and I λ2 represent the radiation intensities measured at wavelengths λ1 and λ2 respectively;

[0022] When inverting the concentration of any gas component in the flame, establish an equation at the characteristic absorption wavelength of the said gas using the Beer-Lambert law:

[0023]

[0024] where τ λ = κ λ CL, C is the gas concentration field, and L is the optical path length; Measure the background radiation I λ0 , and under the given I λ , fit the concentration field C of the said gas using the least squares method:

[0025]

[0026] Furthermore, in the present invention, in step two, before inverting the radiation data of the temperature field and gas concentration field of the flame, there is also a step of correcting the light intensity information in the three-dimensional dataset of the flame field:

[0027] Represent the light intensity information in different polarization states using the Stokes vector:

[0028]

[0029] where I is the total light intensity, Q = I0 - I 90 , U = I 45 - I 135 , the subscript of I represents the polarization angle, and the subscripts I0, I 90 , I 45 , I 135 represent the light intensities at polarization degrees of 0, 90, 45, and 135 degrees respectively, and V is the circular polarization component;

[0030] Differentiate and correct the scattered radiation and direct radiation by measuring the polarization degree P to obtain the corrected light intensity information:

[0031]

[0032] Among them, represents the direct radiation spectral information, represents the flame radiation spectral information measured by the hyperspectral camera.

[0033] Furthermore, in the present invention, in step three, the formulas for establishing the three-dimensional scalar fields of flame temperature and gas concentration are the same, which is:

[0034]

[0035] Among them, f(x,y) is the flame temperature or concentration distribution, p θ is the flame radiation projection data at the angle θ, δ(x cosθ + y sinθ - m) is the Dirac function, ensuring that the projection is only integrated along the straight line of x cosθ + y sinθ = m, m is the projection coordinate, x and y are the spatial position coordinates of the pixel points in the flame, θ is the projection angle, that is, the angle relative to the horizontal axis during camera acquisition, and m is the projection coordinate of the flame image in the θ angle direction.

[0036] Furthermore, in the present invention, in step four, the objective function is:

[0037]

[0038] Among them, F k represents the flame temperature field T k (x,y,z) or the gas concentration field C k (x,y,z) at the k-th iteration, P i0 is the actually measured radiation spectral information and the light intensity information under different polarization states, P i (T k ,C k ) is the temperature or gas concentration field at the k-th iteration.

[0039] Furthermore, in the present invention, in step four, the condition for the objective function to reach the iteration stop is: the absolute value of the difference between the temperature distributions of the flames in two adjacent iterations or the absolute value of the difference between the gas concentration distribution fields of the flames in two adjacent iterations is less than the corresponding threshold;

[0040] max|T k+1 (x,y,z)-T k (x,y,z)|<∈ T

[0041] max|C k+1(x,y,z)-C k (x,y,z)|<∈ C

[0042] where ∈ T and ∈ C are the set convergence thresholds of temperature and gas concentration fields, usually taking 10 -4 ~10 -6 .

[0043] A three-dimensional temperature and concentration detection device for flame combustion based on hyperspectral camera and polarization technology, implemented based on the above method, includes: a hyperspectral camera, a polarization camera, a camera track, a data collector, and a data processing system;

[0044] The camera track is circular. The hyperspectral camera and the polarization camera are fixed together and slidably connected to the camera track through a bracket. The hyperspectral camera and the polarization camera respectively collect the radiation spectrum information of the flame and the light intensity information under different polarization states, and transmit the collected information to the data collector. The data collector performs analog-to-digital conversion on the received signal and then transmits it to the data processing system. The data processing system uses the radiation spectrum information and the light intensity information under different polarization states to establish a three-dimensional scalar field of flame temperature and gas concentration, and calculates the three-dimensional temperature and different gas concentration data of the flame combustion when the temperature distribution T and the concentration distribution C of the flame enter a stable state.

[0045] Using a hyperspectral camera and a polarization camera to jointly measure the three-dimensional temperature field and concentration field of the flame can combine the advantages of both, providing more comprehensive and accurate information from different angles and levels, enabling the comprehensive acquisition of multi-dimensional information and improving the measurement accuracy. The hyperspectral camera can capture multiple spectral bands ranging from visible light to near-infrared, providing rich chemical composition information. The absorption and emission spectral lines of various gases (such as CO, CO2, H2O, etc.) in the flame can accurately reflect the concentrations of different components and the temperature distribution of the flame through the hyperspectral camera. The polarization camera focuses on measuring the polarization characteristics of scattered light and reflected light in the flame. The polarization information helps to exclude the scattered signal in the radiation signal and improve the measurement accuracy. By combining the data of both, multiple physical quantities such as the chemical composition and temperature distribution of the flame can be obtained simultaneously to comprehensively evaluate the characteristics of the flame.

[0046] At the same time, it can overcome the limitations of a single technology. Although the hyperspectral camera can provide rich molecular spectral information, in some complex combustion environments, such as when the flame temperature is extremely high or the optical thickness is large, the spectral signal may be interfered by multiple effects such as absorption and scattering, resulting in an increase in data noise. The polarization camera can effectively reduce this influence because the polarization information is mainly related to the scattering process, thereby improving the flame reconstruction accuracy and reducing the measurement error. Description of the Drawings

[0047] Figure 1 This is the flowchart of the method according to the present invention;

[0048] Figure 2 This is a schematic diagram of the device according to the present invention for collecting flame information;

[0049] Figure 3 This is a schematic diagram of the installation position of the vibrating piece at the front end of the hyperspectral camera. Detailed implementation manners

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts belong to the scope of protection of the present invention. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0051] Detailed implementation manner one: Refer to Figure 1 This detailed implementation manner will be specifically described. The three-dimensional temperature and concentration detection method of flame combustion based on a hyperspectral camera and polarization technology described in this detailed implementation manner includes:

[0052] Step 1: Use a hyperspectral camera and a polarization camera to jointly collect the radiation spectrum information of the flame and the light intensity information under different polarization states from multiple perspectives and positions. Utilize the radiation spectrum information of the flame and the light intensity information under different polarization states to calculate the radiation data in multiple bands, obtain the three-dimensional spatial information of the flame, and form a three-dimensional data set of the flame field;

[0053] Step 2: Based on the radiation transfer theory, utilize the three-dimensional data set of the flame field, and adopt Planck's law and the Beer-Lambert law followed by the gas absorption spectrum to invert the radiation data of the temperature field and gas concentration field of the flame;

[0054] Step 3: Adopt a tomography method, utilize the spectral information of the flame, the light intensity information under different polarization states, and combine the radiation data of the temperature field and gas concentration field of the flame to establish a three-dimensional scalar field of the flame temperature and gas concentration;

[0055] Step 4: Establish an objective function using the three-dimensional scalar fields of flame temperature and gas concentration and the spectral information of the flame or the light intensity information under different polarization states. Adopt an optimization method based on least squares to iteratively optimize the three-dimensional scalar fields of flame temperature and gas concentration in the objective function, and judge in real time whether the objective function reaches the iteration stop condition. If so, obtain the numerical values of the three-dimensional scalar fields of the corresponding flame temperature and gas concentration. Otherwise, return to Step 1 until the objective function reaches the iteration stop condition, and obtain the three-dimensional temperature of flame combustion and the concentration data of different gas components.

[0056] In the present invention, the multi-physical quantity field of spectral radiation characteristics is actually the flame data collected by the hyperspectral camera and the polarization camera, and other physical quantity fields that can be retrieved in addition to the temperature field and the concentration field, such as the velocity field. The present invention can simultaneously retrieve and obtain the multi-physical quantity field of spectral radiation characteristics, while the prior art fails to retrieve these physical quantity fields with high precision, resulting in large errors.

[0057] Furthermore, in the present invention, in Step 2, the specific method for retrieving the radiation data of the temperature field and the gas concentration field of the flame is as follows:

[0058] First, establish the transfer equation:

[0059]

[0060] where I λ is the radiation intensity at wavelength λ, κ λ is the gas absorption coefficient, σ λ is the scattering coefficient, J λ is the scattering source term, s is the path length along the light propagation direction, B λ is the blackbody radiation, and B λ (T) is obtained by Planck's law:

[0061]

[0062] where c1 is the first radiation constant, 3.7419×10 -16 W·m 2 ; c2 is the second radiation constant, 1.4388×10 - 2 m·K; Use the first and second terms of the transfer equation to retrieve the temperature and gas concentration:

[0063]

[0064] Solve for the temperature field T:

[0065]

[0066] where I λ1 and I λ2respectively represent the radiation intensities measured at wavelengths λ1 and λ2;

[0067] When inverting the concentration of any gas component in the flame, an equation is established at the characteristic absorption wavelength of the gas by using the Beer-Lambert law:

[0068]

[0069] where τ λ = κ λ CL, C is the gas concentration field, and L is the optical path length; the background radiation I λ0 is measured through experiments. Given I λ , the concentration field C of the gas is fitted by using the least squares method:

[0070]

[0071] Furthermore, in the present invention, in step two, before inverting the radiation data of the flame temperature field and gas concentration field, there is also a step of correcting the light intensity information in the three-dimensional dataset of the flame field:

[0072] The Stokes vector is used to represent the light intensity information in different polarization states:

[0073]

[0074] where I is the total light intensity, Q = I0 - I 90 , U = I 45 - I 135 , the subscript of I represents the polarization angle, and the subscripts I0, I 90 , I 45 , I 135 respectively represent the light intensities when the polarization degrees are 0, 90, 45, and 135 degrees, and V is the circular polarization component;

[0075] The scattered radiation and direct radiation are distinguished and corrected by measuring the polarization degree P to obtain the corrected light intensity information:

[0076]

[0077] where represents the direct radiation spectral information, represents the flame radiation spectral information measured by the hyperspectral camera.

[0078] Furthermore, in the present invention, in step three, the formulas for establishing the three-dimensional scalar fields of the flame temperature and gas concentration are the same, which is:

[0079]

[0080] Among them, f(x,y) is the flame temperature or concentration distribution, and p θ is the flame radiation projection data at the angle θ, δ(x cosθ + y sinθ - m) is the Dirac function, ensuring that the projection is integrated only along the line of x cosθ + y sinθ = m. Here, m is the projection coordinate, x and y are the spatial position coordinates of the pixel points in the flame, θ is the projection angle, that is, the angle relative to the horizontal axis during camera acquisition, and m is the projection coordinate of the flame image in the θ angle direction.

[0081] Further, in the present invention, in step four, the objective function is:

[0082]

[0083] Among them, F k represents the flame temperature field T k (x,y,z) at the k-th iteration or the gas concentration field C k (x,y,z) at the k-th iteration. P i0 is the actually measured radiation spectrum information and the light intensity information under different polarization states. P i (T k ,C k ) is the radiation spectrum information and the light intensity information under different polarization states obtained from the temperature and gas concentration fields at the k-th iteration.

[0084] Further, in the present invention, in step four, the condition for the objective function to reach the iteration stop condition is: determine whether the absolute value of the difference between the temperature distributions of the flames in two adjacent iterations or the absolute value of the difference between the gas concentration distribution fields of the flames in two adjacent iterations is less than the corresponding threshold:

[0085] max|T k+1 (x,y,z) - T k (x,y,z)| < ∈ T

[0086] max|C k+1 (x,y,z) - C k (x,y,z)| < ∈ C

[0087] Among them, ∈ T and ∈ C are the set convergence thresholds for the temperature and gas concentration fields, usually taking values from 10 -4 to 10 -6 .

[0088] Specific Embodiment 2: Refer to Figure 2Specifically describing this embodiment, the three-dimensional temperature and concentration detection device for flame combustion based on hyperspectral camera and polarization technology is implemented based on the three-dimensional temperature and concentration detection method for flame combustion based on hyperspectral camera and polarization technology described in the first specific embodiment, and includes: a hyperspectral camera 1, a polarization camera 2, a camera track 3, a data collector 4, and a data processing system 5;

[0089] The camera track 3 is circular. The hyperspectral camera 1 and the polarization camera 2 are fixed together and are slidably connected to the camera track 3 through a bracket. The hyperspectral camera 1 and the polarization camera 2 respectively collect the radiation spectrum information of the flame and the light intensity information under different polarization states, and transmit the collected information to the data collector 4. The data collector 4 performs analog-to-digital conversion on the received signal and then transmits it to the data processing system 5. The data processing system 5 uses the radiation spectrum information and the light intensity information under different polarization states to establish a three-dimensional scalar field of the flame temperature and gas concentration, and calculates the three-dimensional temperature and different gas concentration data of the flame combustion when the temperature distribution T and the concentration distribution C of the flame enter a stable state.

[0090] In order to solve the technical problems that the spatial resolution of the single detection method based on the spectrum is low, the error of the collected radiation signal is large, and the high-precision cooperative measurement of multiple physical quantity fields such as temperature, gas component concentration, and spectral radiation characteristics cannot be realized, the present invention proposes the detection of the three-dimensional temperature field and concentration field of flame combustion based on hyperspectral camera and polarization technology.

[0091] To achieve the above object, the following technical solutions are adopted:

[0092] Step 1: Obtain flame radiation information:

[0093] Use a hyperspectral camera and a polarization camera to obtain multi-band radiation data, and focus on collecting the absorption and emission spectral lines containing temperature and concentration radiation lines, such as the absorption and emission spectral lines of molecules such as CO, CO2, and H2O. Select multiple viewing angles and positions, and ensure that the fields of view of the two cameras are consistent to synchronously obtain the three-dimensional spatial information of the flame and form a three-dimensional data set. The hyperspectral and polarization cameras need to work synchronously to ensure that the data at the same time point can be accurately fused. Select a hyperspectral camera (covering 400nm to 2500nm); a polarization camera that can measure multiple polarization angles (such as 0°, 45°, 90°, 135°). Adopt a synchronous trigger system to realize the time-synchronous measurement of the cameras.

[0094] Step 2: Invert the temperature field and concentration field:

[0095] The inversion of the temperature field and concentration field of the flame is mainly based on the radiation transfer theory, combined with the spectral radiation information measured by hyperspectral measurement and the scattering correction information measured by polarization measurement, to achieve more accurate inversion of temperature and concentration. The light radiation in the flame is affected by the absorption, scattering of gas and soot particles, and its own radiation. Its transfer equation can be expressed as

[0096]

[0097] where I λ is the radiation intensity at wavelength λ, κ λ is the gas absorption coefficient, which is determined by temperature and concentration. σ λ is the scattering coefficient, J λ is the scattering source term, s is the path length along the light propagation direction, B λ is the blackbody radiation, which satisfies Planck's law:

[0098]

[0099] where c1 = 2hc 2 , c2 = hc / k B are the first and second radiation constants.

[0100] For hyperspectral measurement, the first and second terms are mainly used to invert temperature and concentration. The hyperspectral camera is mainly based on the spectral radiation characteristics of the flame. By fitting the radiation intensity I λ and the blackbody radiation B λ the temperature is inverted. Assume that the radiation intensities measured at two wavelengths are I λ1 and I λ2 respectively. According to Planck's law, the temperature can be deduced:

[0101]

[0102] The temperature field T can be solved by this formula:

[0103]

[0104] To invert the concentration field of the flame through radiation intensity data, the gases in the flame (such as CO, CO2, H2O) have characteristic absorption spectra at specific wavelengths. The hyperspectral camera can measure the spectral radiation of the flame in multiple wavelength ranges and invert the concentration field of the gas by analyzing these absorption characteristics. The gas absorption spectrum follows the Beer-Lambert law, the exponential decay relationship of light intensity with optical path:

[0105] Beer-Lambert law:

[0106]

[0107] where Iλ0 is the background radiation intensity when there is no flame, and τ λ is the optical thickness, which describes the absorption of gas at a specific wavelength.

[0108] From the radiation transfer equation, in the case of neglecting scattering, the transmission formula of the flame layer can be obtained by integration:

[0109]

[0110] where is the transmittance of the flame. The first term is the intensity of the background light passing through the flame, which is affected by gas absorption. The second term is the radiation contribution of the flame gas itself.

[0111] When retrieving a certain gas (such as CO2) in the flame, an equation can be established at its characteristic absorption wavelengths (2.0μm, 4.3μm):

[0112]

[0113] where τ λ = κ λ CL, C is the gas concentration field, and L is the optical path length. By experimentally measuring the background radiation I λ0 , in the case of a given I λ , the least squares method can be used to fit the gas concentration field C:

[0114]

[0115] However, due to the scattering effects of soot particles and aerosols in the flame, the directly measured spectral data may be affected by deviations, thereby reducing the inversion accuracy of temperature and concentration. The polarization camera can effectively separate the direct radiation and the scattered radiation, thus correcting the measurement results of the hyperspectral camera and improving the inversion accuracy.

[0116] The polarized light can be represented by the Stokes vector:

[0117]

[0118] Ensure that the fields of view of the two cameras are consistent and synchronously obtain the flame radiation information. Among them, I is the total light intensity, Q = I0 - I 90 , U = I 45 - I 135 , and the subscript represents the polarization angle. V is the circular polarization component.

[0119] The degree of polarization is used to characterize the contribution of scattered light. Scattering in the flame is mainly caused by Mie scattering or Rayleigh scattering. The scattered light has a higher degree of polarization than direct radiation, and this property can be used to separate the two. A polarization camera can distinguish scattered radiation from direct radiation by measuring the degree of polarization P.

[0120]

[0121] The light intensity measured by the hyperspectral camera is composed of direct radiation and scattered radiation :

[0122]

[0123] If the degree of polarization of the scattered light is P and the degree of polarization of the direct radiation is close to zero, then:

[0124]

[0125] Therefore, the direct radiation component can be corrected to:

[0126]

[0127] The corrected radiation intensity data can be used for the inversion of the temperature field and concentration field, thereby improving the reconstruction accuracy.

[0128] Step 3: Reconstruct the three-dimensional temperature field and concentration field of the flame. Using the tomography method, based on hyperspectral and polarization data at multiple angles, reconstruct the three-dimensional temperature field and concentration field of the flame. The temperature and concentration fields of the flame can be regarded as three-dimensional scalar fields, and reconstruction is carried out using projection data:

[0129]

[0130] where f(x, y) is the temperature or concentration distribution of the flame, p θ is the projection data at angle θ, δ(x cosθ + y sinθ - m) is the Dirac function, ensuring that the projection is integrated only along the line x cosθ + y sinθ = m, and m is the projection coordinate.

[0131] Step 4: Convergence judgment. To determine whether the inversion results of the flame temperature field and concentration field converge, we define the objective function F k , which represents the error measure of the current solution. In the inversion method based on least squares optimization, the objective function is usually expressed as:

[0132]

[0133] where P i0 is the actually measured hyperspectral or polarization projection data, Pi (T k , C k ) are the projection data calculated from the temperature and concentration fields obtained by inversion at the k-th iteration. To make the objective function converge, F k should be as small as possible. It is judged whether the change in the objective function between two adjacent iterations is less than the set threshold ∈. If it is less, it is considered to have converged:

[0134]

[0135] In the iterative process of the flame temperature field and concentration field, if the change in the objective function at each point is less than ∈, it is considered that the solution has converged.

[0136] max|T k+1 (x, y, z) - T k (x, y, z)| < ∈ T

[0137] max|C k+1 (x, y, z) - C k (x, y, z)| < ∈ C

[0138] where ∈ T and ∈ C are the set convergence thresholds for the temperature and concentration fields, usually taking 10 -4 ~10 -6 . The convergence of the objective function indicates that the temperature distribution T and concentration distribution C have stabilized and proceed to the next step. If not converged, return to step one and continue to optimize the solution process.

[0139] Step Five: Data fusion optimization. To improve the reconstruction accuracy, during the tomographic reconstruction process, the data after polarization correction is used to reduce the influence of scattering on the reconstruction accuracy. At the same time, Tikhonov regularization can also be used to suppress noise.

[0140] Step Six: Error analysis and compensation. During the reconstruction process, both tomographic imaging and data fusion will introduce errors. For example, the hyperspectral camera and polarization camera need to be spatially registered. If the registration is inaccurate, it will affect the fusion accuracy; at the same time, the flame is dynamically changing. If the synchronization error between the two cameras is large, it may lead to inconsistent data; in actual measurements, the obtained projection data may be limited, resulting in insufficient information during inversion.

[0141] The corresponding optimization methods include optimizing multi-sensor matching, using calibration points to perform spatial registration of the hyperspectral camera and polarization camera to improve the fusion accuracy. At the same time, optimize the time synchronization, use a high-speed synchronous trigger system to ensure that the two cameras are exposed at the same time. It is also possible to try to remove the camera noise and reduce the sensor noise using dark field correction and flat field correction.

[0142] Step 7: Output the final result. After completing all calculations, output the final combustion flame parameters, including the temperature field T, gas concentration field C, and spectral radiation characteristics k.

[0143] In the present invention, the hyperspectral camera and the polarization camera are arranged side by side. A total of three groups of cameras are set up in the form of two cameras in a group. Each group of cameras is evenly placed on a circular orbit, and the angle between each group is 120°. The orbit is designed such that the camera array can rotate 360° to achieve omnidirectional flame measurement. The hyperspectral camera is mainly responsible for collecting the spectral information of the flame, covering the visible to near-infrared band (about 400 nm to 2500 nm), for reconstructing the temperature field of the flame and the concentration fields of water, CO2, etc. At the same time, the polarization camera mainly collects the light intensity information under different polarization states in the flame. These data can be used to correct the errors caused by scattered radiation and improve the reconstruction accuracy. The synchronous operation of the hyperspectral camera and the polarization camera enables the acquisition data of the two to complement each other, providing omnidirectional information of the flame, thereby comprehensively and highly accurately describing the multi-physical field characteristics of the flame.

[0144] The present invention also provides a three-dimensional temperature and concentration detection device for flame combustion based on a hyperspectral camera and polarization technology, specifically as Figure 2 shown. The device is used to implement the three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology; the device includes: a hyperspectral camera 1, a polarization camera 2, a camera orbit 3, a data collector 4, and a data processing system 5; the hyperspectral camera 1 and the polarization camera 2 are placed on a rotatable orbit, capable of photographing the flame omnidirectionally, and transmitting their spontaneous radiation information to the data collector 4 through the output end; the data collector 4 transmits the detected laser absorption spectral information to the data processing system 5 through the output end; the data processing system 5 processes the obtained hyperspectral signal and polarization signal, and reconstructs the high-spatial-resolution and high-precision temperature, component concentration, and spectral radiation characteristic field distribution of the combustion flame through tomographic measurement methods.

[0145] The present invention can also integrate the hyperspectral camera with the polarizer 1.1 to achieve the function of fixing the two cameras, as Figure 3As shown, by adjusting the angle of the polarizer, the radiation information of the flame at different polarization angles is obtained. The polarizer A is installed in front of the hyperspectral camera B and can rotate at different polarization angles (such as 0°, 45°, 90°, 135°, etc.). Each time a shot is taken, the angle of the polarizer needs to be adjusted in a preset order, and the spectral data of the corresponding band is taken. Through multi-angle polarization acquisition, the light intensity information of the flame in different polarization states is recorded. These data are used to invert the temperature and concentration distributions in the flame. At each polarization angle, the hyperspectral camera will collect the corresponding spectral data (covering the visible to near-infrared band, approximately 400 nm to 2500 nm). Since different polarization angles will affect the intensity of the flame radiation, by comparing the spectral information of different polarization angles, the measurement error caused by scattered radiation to the flame's own radiation can be excluded.

[0146] Although the present invention has been described herein with reference to particular embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be designed, provided that they do not depart from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and the features described herein may be combined in ways different from those described in the original claims. It should also be understood that the features described in connection with a single embodiment may be used in other described embodiments.

Claims

1. A three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology, characterized in that, Including: Step 1: Use a hyperspectral camera and a polarization camera to jointly collect the radiation spectral information of the flame and the light intensity information under different polarization states from multiple perspectives and positions. Utilize the collected information to calculate the radiation data of multiple bands of the flame, obtain the three-dimensional spatial information of the flame, and form a three-dimensional dataset of the flame field; Step 2: Based on the radiation transfer theory, use the three-dimensional dataset of the flame field, and adopt Planck's law and the Beer-Lambert law followed by the gas absorption spectrum to invert the radiation data of the temperature field and gas concentration field of the flame; Step 3: Adopt a tomography method, utilize the spectral information of the flame, the light intensity information under different polarization states, combined with the radiation data of the temperature field and gas concentration field of the flame, to establish a three-dimensional scalar field of the flame temperature and gas concentration; Step 4: Establish an objective function using the three-dimensional scalar field of the flame temperature and gas concentration and the spectral information of the flame or the light intensity information under different polarization states. Adopt an optimization method based on the least squares to iteratively optimize the three-dimensional scalar field of the flame temperature and gas concentration in the objective function, and judge in real time whether the objective function reaches the iteration stop condition. If so, obtain the corresponding three-dimensional temperature of the flame combustion and the concentration data of different gas components. Otherwise, return to Step 1 until the objective function reaches the iteration stop condition, and obtain the three-dimensional temperature of the flame combustion and the concentration data of different gas components.

2. The three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology according to claim 1, wherein In Step 2, the specific method for inverting the radiation data of the temperature field and gas concentration field of the flame is as follows: First, establish a transfer equation: where, I λ is the radiation intensity at wavelength λ, κ λ is the gas absorption coefficient, σ λ is the scattering coefficient, J λ is the scattering source term, s is the path length along the light propagation direction, B λ is the blackbody radiation, and B λ (T) is obtained using Planck's law: where, c1 is the first radiation constant, 3.7419×10 -16 W·m 2 ; c2 is the second radiation constant, 1.4388×10 -2 m·K; Invert the temperature and gas concentration using the first and second terms of the transfer equation: Solve for the temperature field T: where I λ1 and I λ2 represent the radiation intensities measured at wavelengths λ1 and λ2, respectively; When inverting the concentration of any gas component in the flame, establish an equation at the characteristic absorption wavelength of the gas using the Beer-Lambert law: where τ λ = κ λ CL, C is the gas concentration field, L is the optical path length; the background radiation I λ0 is measured through experiments. Given I λ , the concentration field C of the said gas is fitted by the least squares method:

3. The three-dimensional temperature and concentration detection method for flame combustion based on hyperspectral cameras and polarization technology according to claim 2, characterized in that, In Step 2, before inverting the radiation data of the temperature field and gas concentration field of the flame, it also includes a step of correcting the light intensity information in the three-dimensional dataset of the flame field: Represent the light intensity information under different polarization states using the Stokes vector: Wherein, I is the total light intensity, Q = I0 - I 90 , U = I 45 -I 135 , the subscript of I represents the polarization angle, and the subscripts I0, I 90 , I 45 , I 135 respectively represent the light intensities when the degrees of polarization are 0, 90, 45, and 135 degrees, and V is the circular polarization component; Distinguish and correct the scattered radiation and direct radiation by measuring the degree of polarization P to obtain the corrected light intensity information: Among them, represents the direct radiation spectral information, represents the flame radiation spectral information measured by the hyperspectral camera.

4. The three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology according to claim 3, characterized in that, In Step 3, the formulas for the three-dimensional scalar fields of the flame temperature and gas concentration are the same, which is: Among them, f(x, y) is the flame temperature or concentration distribution, and p θ is the flame radiation projection data at the angle θ, x and y are the spatial position coordinates of the pixel points in the flame, δ(xcosθ + ysinθ - m) is the Dirac function, θ is the projection angle, that is, the angle relative to the horizontal axis during camera acquisition, and m is the projection coordinate of the flame image in the θ angle direction.

5. The three-dimensional temperature and concentration detection method of flame combustion based on hyperspectral camera and polarization technology according to claim 4, characterized in that In Step 4, the objective function is: Among them, F k represents the flame temperature field T k (x, y, z) at the k-th iteration or the gas concentration field C k (x, y, z), and P i0 is the radiation spectrum information measured actually and the light intensity information under different polarization states. P i (T k , C k ) is the temperature or gas concentration field at the k-th iteration.

6. The three-dimensional temperature and concentration detection method for flame combustion based on hyperspectral camera and polarization technology according to claim 5, characterized in that, In Step 4, the condition for the objective function to reach the iteration stop condition is: the absolute value of the difference in the temperature distribution of the flame between two adjacent iterations or the absolute value of the difference in the gas concentration distribution field of the flame between two adjacent iterations is less than the corresponding threshold; where, ∈ T and ∈ C are the set convergence thresholds for temperature and gas concentration fields, usually taking 10 -4 ~10 -6 .

7. A three-dimensional temperature and concentration detection device for flame combustion based on a hyperspectral camera and polarization technology, which is implemented based on the three-dimensional temperature and concentration detection method for flame combustion based on a hyperspectral camera and polarization technology according to any one of claims 1-6, characterized in that Including: A hyperspectral camera (1), a polarization camera (2), a camera track (3), a data collector (4), and a data processing system (5); The camera track (3) is circular. The hyperspectral camera (1) and the polarization camera (2) are fixed together and are slidably connected to the camera track (3) through a bracket. The hyperspectral camera (1) and the polarization camera (2) respectively collect the radiation spectrum information of the flame and the light intensity information under different polarization states, and transmit the collected information to the data collector (4). The data collector (4) performs analog-to-digital conversion on the received signal and then transmits it to the data processing system (5). The data processing system (5) uses the radiation spectrum information and the light intensity information under different polarization states to establish a three-dimensional scalar field of the flame temperature and gas concentration, and calculates the three-dimensional temperature and different gas concentration data of the flame combustion when the temperature distribution and concentration distribution of the flame enter a stable state.

8. The three-dimensional temperature and concentration detection device for flame combustion based on a hyperspectral camera and polarization technology according to claim 7, characterized in that The number of both the hyperspectral camera (1) and the polarization camera (2) is three.

Images