A method and system for online detection of coal quality entering the furnace

Abstract

This invention relates to the field of coal quality testing technology, specifically to an online method and system for detecting coal quality entering the furnace; it includes: a multimodal data acquisition step: acquiring microwave scattering parameters and surface element characteristic spectra using a microwave tomography sensor array and a surface spectrometer, respectively; a physical field inversion and modeling step: reconstructing the complex permittivity distribution map based on the scattering parameters to establish a three-dimensional physical distribution model including density and moisture; a field-spectrum coupling deduction step: using the characteristic spectra as boundary constraints, and combining them with the complex permittivity distribution map to deduce the internal chemical element distribution; and a volume-weighted quantization step: performing volume-weighted integration on the three-dimensional physical model and chemical distribution to output comprehensive coal quality parameters; this invention eliminates the detection blind spots of single technologies and significantly improves the representativeness and safety of full-section detection of coal entering the furnace.

Description

Technical Field

[0001] This invention relates to the field of coal quality testing technology, specifically to an online method and system for testing the quality of coal entering the furnace. Background Technology

[0002] In the current coal conveying scenario of thermal power generation, the coal flow into the furnace exhibits physical characteristics of large flow rate, high-speed transmission, and highly heterogeneous internal composition; in order to ensure combustion efficiency and unit safety, it is necessary to conduct real-time online detection of the coal quality entering the furnace.

[0003] Existing detection methods generally employ single-mode techniques, which present significant technical bottlenecks: traditional X-ray transmission techniques, while possessing penetrating capabilities, pose radiation safety hazards and cannot resolve specific chemical elemental compositions; while optical detection methods such as laser-induced breakdown spectroscopy, although capable of obtaining high-precision elemental spectra, are limited to surface detection of coal flows and cannot penetrate deep coal bodies; given the spatial differences in coal density distribution and moisture content, high-ash gangue and high-moisture coal masses often exhibit similar physical characteristics, leading to severe representativeness deficiencies in single-density assumptions or surface data averaging; simultaneously, moisture fluctuations can produce a nonlinear quenching effect on spectral signals, directly affecting the accuracy of elemental inversion.

[0004] Therefore, how to effectively integrate deep physical structure and surface chemical information without using a radioactive source, eliminate moisture interference, and solve the detection blind spots caused by physical ambiguity, so as to improve the representativeness and accuracy of full-section coal quality testing, has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide an online detection method and system for coal entering the furnace, aiming to resolve the technical contradiction between the inability of high-penetration detection methods to resolve chemical elements and the inability of high-precision spectroscopic methods to penetrate deep coal layers in existing technologies. This invention eliminates the detection blind spots of single technologies without using a radioactive source, significantly improving the representativeness and safety of full-section coal quality detection of coal entering the furnace. Specifically, the technical solution of this invention is as follows:

[0006] A method for online detection of coal quality entering the furnace includes the following steps:

[0007] The microwave tomography sensor array and surface spectral detector were used to obtain the multi-angle microwave scattering parameters and surface element characteristic spectra of the coal flow to be tested.

[0008] Based on the multi-angle microwave scattering parameters of the micro-element, nonlinear physical field inversion calculation is performed to reconstruct the complex permittivity distribution map of the cross-section of the coal flow to be tested in the furnace, and a three-dimensional physical distribution model including density distribution and moisture distribution is established.

[0009] Using the surface elemental characteristic spectrum of the micro-element as a boundary constraint, and combined with the complex permittivity distribution map of the micro-element, the chemical element distribution inside the coal flow to be tested is deduced through the field-spectrum mapping model.

[0010] Based on the established three-dimensional physical distribution model of the micro-element and the deduced chemical element distribution of the micro-element, a volume-weighted integral algorithm is used for coal quality analysis, and comprehensive coal quality parameters are output.

[0011] Preferably, the multi-angle microwave scattering parameters and surface elemental characteristic spectra of the coal flow to be tested are acquired using a microwave tomography sensor array and a surface spectrometer, including:

[0012] The microwave tomography sensor array surrounding the coal conveying channel is controlled to transmit and receive microwave signals that penetrate the coal flow to be tested in sequence according to a preset time sequence, and the scattering parameter matrix under different projection angles is recorded.

[0013] A laser-induced breakdown spectroscopy probe positioned above the coal flow to be tested is controlled to acquire the plasma emission spectrum of the surface layer of the coal flow at high frequency, generating a surface elemental characteristic spectrum containing carbon, hydrogen, and sulfur characteristic peaks.

[0014] Preferably, based on the multi-angle microwave scattering parameters of the micro-element, a nonlinear physical field inversion operation is performed to reconstruct the complex permittivity distribution map of the cross-section of the coal flow to be measured, including:

[0015] The microwave scattering parameters of the micro-element from multiple angles are calibrated and denoised to generate standard scattering data.

[0016] The standard scattering data of the micro-element is input into a preset nonlinear tomographic imaging algorithm model, and the residual between the measured value and the calculated value is minimized through iterative calculation.

[0017] When the residual of the infinitesimal element converges to the preset threshold range, the complex permittivity distribution map of the current iteration step is output, where the real part of the complex permittivity represents the coal flow density and the imaginary part represents the local moisture content.

[0018] Preferably, the surface elemental characteristic spectrum of the micro-element is used as a boundary constraint condition, combined with the complex permittivity distribution map of the micro-element, and the chemical elemental distribution inside the coal flow to be tested is deduced through a field-spectrum mapping model, including:

[0019] Local moisture values ​​of the surface region were extracted from the complex permittivity distribution map of the micro-element;

[0020] By using the local moisture value of the micro-element, matrix effect compensation is performed on the surface elemental characteristic spectrum of the micro-element at the same spatial location to eliminate the nonlinear interference of moisture fluctuation on spectral intensity and generate corrected elemental concentration data.

[0021] Using the elemental concentration data corrected by the micro-element as boundary values, a semi-supervised manifold learning algorithm is used to map the surface chemical information to the internal volume region of the coal flow to be tested along the density gradient direction of the three-dimensional physical distribution model of the micro-element.

[0022] Preferably, based on the established three-dimensional physical distribution model of the micro-element and the deduced chemical element distribution of the micro-element, a volume-weighted integral algorithm is used for coal quality analysis, including:

[0023] The coal flow to be measured in the micro-element is divided into multiple micro-elements;

[0024] The density weight of each infinitesimal element is determined based on the distribution diagram of the complex permittivity of the infinitesimal element.

[0025] The elemental concentration of each micro-element is determined based on the chemical elemental distribution of the micro-element.

[0026] Density-weighted volume integrals are performed on the elemental component concentrations of all micro-elements to calculate the weighted average calorific value, total sulfur content, and total ash content of the entire coal flow to be tested.

[0027] Preferred options also include:

[0028] Real-time monitoring of abrupt changes in dielectric constant in the distribution diagram of complex dielectric constant of infinitesimal elements;

[0029] Among them, the preset metal threshold is greater than the preset rock threshold;

[0030] If the dielectric constant of the abrupt change region of the micro-element is greater than the metal threshold of the micro-element, it is determined that there is a metal foreign object in the coal flow to be tested and an rejection indication signal is generated.

[0031] If the dielectric constant of the abrupt change region of the micro-element is less than or equal to the metal threshold of the micro-element and greater than the rock threshold of the micro-element, it is determined that there are large pieces of gangue in the coal flow to be tested and a warning signal is generated.

[0032] If the dielectric constant of the abrupt change region of the micro-element is less than or equal to the rock threshold of the micro-element, it is determined to be normal coal quality, and the micro-element coal quality analysis step is continued.

[0033] Preferably, after outputting the comprehensive coal quality parameters from the micro-element, the following is also included:

[0034] The comprehensive coal quality parameters of the micro-element were compared with the preset combustion control model;

[0035] If the weighted average calorific value in the comprehensive coal quality parameters of the micro-element is lower than the preset combustion efficiency threshold, an adjustment feedback signal for the coal mill speed or coal feed rate will be generated.

[0036] The micro-element adjustment feedback signal is transmitted to the boiler distributed control system to achieve closed-loop combustion control.

[0037] Preferably, matrix effect compensation is performed using the local moisture value of a micro-element to analyze the surface elemental characteristic spectrum of micro-elements at the same spatial location, including:

[0038] A moisture-spectral intensity attenuation database was constructed, and spectral correction coefficient curves under different moisture contents were obtained by fitting.

[0039] Based on the local moisture value of the micro-element, find the corresponding target correction coefficient in the spectral correction coefficient curve of the micro-element;

[0040] Multiply the original intensity value of the elemental characteristic spectrum on the surface of the micro-element by the target correction factor of the micro-element to obtain the corrected spectrum that reflects the true elemental concentration.

[0041] An online coal quality detection system for furnace feed includes:

[0042] The data acquisition unit is used to acquire the multi-angle microwave scattering parameters and surface element characteristic spectrum of the coal flow to be tested through a microwave tomography sensor array and a surface spectral detector, respectively.

[0043] The physical field reconstruction unit is used to perform nonlinear physical field inversion calculations based on the multi-angle microwave scattering parameters of the micro-element, reconstruct the complex permittivity distribution map of the cross section of the coal flow to be measured, and establish a three-dimensional physical distribution model including density distribution and moisture distribution.

[0044] The field-spectrum coupling analysis unit is used to take the surface elemental characteristic spectrum of the micro-element as the boundary constraint condition, and combine it with the complex permittivity distribution map of the micro-element to deduce the chemical element distribution inside the coal flow to be tested through the field-spectrum mapping model.

[0045] The comprehensive parameter calculation unit is used to perform coal quality analysis using a volume-weighted integral algorithm based on the established three-dimensional physical distribution model of the micro-element and the deduced chemical element distribution of the micro-element, and output comprehensive coal quality parameters.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1. This invention, by coupling microwave tomography and surface spectral detection technologies, uses microwaves to reconstruct the internal physical framework and spectroscopy to construct the surface chemical skin, thus resolving the contradiction between the inability of high-penetration methods to resolve elements and the inability of high-precision spectra to penetrate deep layers. This holographic detection mode, which goes from the surface to the inside, can eliminate the detection blind spots of single technologies without using a radioactive source, and significantly improves the representativeness and safety of full-section detection of coal entering the furnace.

[0048] 2. This invention utilizes the local moisture value obtained by microwave inversion to perform matrix effect compensation on spectral data, effectively eliminating the nonlinear interference of moisture fluctuations on plasma emission intensity. This mechanism overcomes the problem of spectral signal distortion caused by moisture quenching effect in wet coal scenarios, ensuring the accuracy of converting surface spectra into real elemental concentration data, thereby providing high-confidence boundary conditions for subsequent internal chemical deduction.

[0049] 3. This invention adopts a field spectrum mapping model based on semi-supervised manifold learning, taking density and moisture distribution as joint physical constraints to drive the surface chemical information to be mapped to the interior along the density gradient. This method abandons the assumption of single density homogeneity and effectively solves the problem of physical ambiguity between high ash gangue and high moisture coal lumps with similar densities but different chemical properties, realizing accurate deduction from two-dimensional surface boundary to three-dimensional interior space.

[0050] 4. This invention uses a volume-weighted integral algorithm to replace the traditional arithmetic mean. It calculates the comprehensive coal quality parameters based on the actual density weight of the micro-element, accurately reflecting the impact of changes in the internal packing density of the coal flow on the contribution of calorific value, thus improving the calculation accuracy. At the same time, by comparing the detection results with the combustion control model and generating feedback signals, it realizes the leap from passive detection to active closed-loop control, effectively suppressing coal quality fluctuations and ensuring boiler combustion efficiency. Attached Figure Description

[0051] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0052] Figure 1 This is a flowchart of the method of the present invention;

[0053] Figure 2 This is a structural diagram of the system of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0055] Example 1:

[0056] Please see Figure 1 An online detection method for coal quality entering the furnace includes the following steps: acquiring multi-angle microwave scattering parameters and surface element characteristic spectra of the coal flow to be tested through a microwave tomography sensor array and a surface spectral detector, respectively;

[0057] Based on multi-angle microwave scattering parameters, nonlinear physical field inversion calculations are performed to reconstruct the complex permittivity distribution map of the cross-section of the coal flow to be tested, and a three-dimensional physical distribution model including density distribution and moisture distribution is established. The surface element characteristic spectrum map is used as a boundary constraint condition, and combined with the complex permittivity distribution map, the chemical element distribution inside the coal flow to be tested is deduced through the field-spectrum mapping model.

[0058] Based on the established three-dimensional physical distribution model and the deduced chemical element distribution, a volume-weighted integral algorithm is used to perform coal quality analysis and output comprehensive coal quality parameters.

[0059] This embodiment details the specific execution logic and core architecture of the above method, which aims to resolve the technical contradiction between the inability of high-penetration detection methods to analyze chemical elements and the inability of high-precision spectroscopic methods to penetrate deep coal bodies in the prior art.

[0060] The system performs a multimodal data concurrent acquisition step, using a microwave tomography sensor array and a surface spectroscopy detector installed in the coal conveyor corridor to acquire multi-angle microwave scattering parameters of the coal flow to be tested. Surface elemental characteristic spectrum Among them, microwave scattering parameters reflect the amplitude attenuation and phase delay information of microwave signals when penetrating non-uniform coal flow, and contain the physical characteristics of the internal structure of coal flow; surface element characteristic spectrum reflects the atomic emission intensity of elements such as carbon, hydrogen, and sulfur in the micro-region of the coal flow surface.

[0061] The system performs nonlinear physical field inversion and three-dimensional model establishment steps. Based on multi-angle microwave scattering parameters, it solves the electromagnetic inverse scattering problem and reconstructs the complex permittivity distribution map of the cross-section of the coal flow entering the furnace. Based on this, it establishes a model including density distribution. With moisture distribution A three-dimensional physical distribution model; this model constitutes the volumetric skeleton of the coal flow, solving the problem of lack of representativeness in traditional spectroscopic techniques;

[0062] The system executes the field-spectrum coupling and chemical distribution deduction steps, using the surface element characteristic spectrum as the boundary constraint condition, combined with the complex permittivity distribution map, and deduces the chemical element distribution inside the coal flow to be tested through the field-spectrum mapping model; in this process, the field-spectrum mapping model uses the three-dimensional physical field reconstructed by microwave, including density and moisture, as the spatial index skeleton, based on the construction principle of physical-chemical multidimensional correlation.

[0063] To address the physical ambiguity in highly heterogeneous mixtures like coal, where high-ash gangue and high-moisture coal lumps may have similar densities but drastically different chemical properties, the model abandons the single assumption of density homogeneity. Instead, it defines the inputs as surface spectral concentration and internal micro-element concentration, which follow a density-moisture joint feature similarity propagation law. That is, it assumes that under the dual physical constraints of density and moisture, adjacent connected regions in the feature space have homogeneous chemical properties, thereby establishing a physically unique derivation path from the two-dimensional surface boundary to the three-dimensional internal space.

[0064] This step specifically employs a density-based manifold learning logic. It utilizes the internal density gradient information provided by the microwave field to construct a weighted adjacency graph in three-dimensional space, defining that the propagation probability between adjacent micro-elements is negatively correlated with the density difference. The known chemical information on the surface is used as the source node on the graph. By solving the graph Laplace equation to minimize the energy function of the entire graph, the chemical information is driven to be mapped and diffused into the internal manifold space along the density isosurface or smooth gradient path, thereby obtaining the estimated value of the element concentration in the internal invisible region.

[0065] The system performs volume-weighted integration and parameter output steps. Based on the established three-dimensional physical distribution model and the deduced chemical element distribution, the volume-weighted integration algorithm is used to perform coal quality analysis and output comprehensive coal quality parameters, such as received basis lower heating value, total sulfur content and total ash content.

[0066] This embodiment constructs a field-spectrum coupled detection architecture, using microwave fields to build the physical skeleton inside the coal flow and surface spectra to build a chemical skin. A specific mapping algorithm is then used to transmit the skin information along the skeleton to the interior, thereby achieving full-volume coal quality reconstruction. This holographic detection mode, from the surface to the interior, can eliminate the detection blind spots of a single technology in the scenario of coal conveyor belts in thermal power plants without using radioactive sources, significantly improving the representativeness and safety of full-section coal quality detection of coal entering the furnace.

[0067] Example 2:

[0068] The microwave tomography sensor array and surface spectral detector are used to acquire the multi-angle microwave scattering parameters and surface element characteristic spectrum of the coal flow to be tested. This includes controlling the microwave tomography sensor array surrounding the coal conveying channel to transmit and receive microwave signals that penetrate the coal flow to be tested in sequence according to a preset time sequence, and recording the scattering parameter matrix under different projection angles.

[0069] A laser-induced breakdown spectroscopy probe positioned above the coal flow to be tested is controlled to acquire the plasma emission spectrum of the surface layer of the coal flow at high frequency, generating a surface elemental characteristic spectrum containing carbon, hydrogen, and sulfur characteristic peaks.

[0070] This embodiment further defines the specific hardware control logic for the data acquisition steps; for the timing acquisition of microwave data, the microwave tomography sensor array is composed of... The system consists of transceiver antennas evenly distributed along the outside of the coal conveying channel; the operating frequency of the microwave signal is set. The background wavenumber is 2.45 GHz in the ISM band. The frequency band was chosen to balance the meter-level penetration depth of electromagnetic waves in coal seams with centimeter-level imaging resolution.

[0071] The control unit controls the antenna to operate according to a preset polling excitation mode: at time... Selected One antenna transmits microwave signals, the rest... One antenna is used as the receiving end to collect the microwave signal after penetrating the coal flow entering the furnace; all antennas are used as transmitting ends to record the scattering parameter matrix under different projection angles. This matrix fully records the electromagnetic response of the coal flow at various angles, providing a complete dataset for subsequent tomographic imaging.

[0072] Simultaneously, for the high-frequency acquisition of spectral data, the laser-induced breakdown spectral probe located directly above the coal flow to be tested is controlled to emit high-energy pulsed lasers onto the surface of the coal flow; the instantaneous high temperature generated by laser focusing causes trace amounts of coal powder to gasify and form plasma; the spectrometer inside the probe acquires the atomic spectrum emitted during the plasma cooling process at a frequency of 10Hz to 50Hz, generating a surface elemental characteristic spectrum containing characteristic peaks of carbon, hydrogen, sulfur and other elements.

[0073] Based on this, in order to solve the spatial alignment problem between the microwave tomography cross section and the spectral detection point on the high-speed flowing coal, the system executes spatiotemporal synchronization logic based on velocity compensation; the physical distance between the central cross section of the microwave sensor array and the breakdown point of the spectral probe is set as... This distance is along the direction of belt travel, and the operating speed of the coal conveyor belt is obtained in real time. ;

[0074] For time Spectral data of the collection The corresponding microwave scattering parameter matrix should be indexed to the time step. Nearby historical data frames; the system constructs a circular buffer in memory to temporarily store microwave scattering data. Only after a successful delay match are the two sets of data packaged and sent to the subsequent processing unit, thereby ensuring spatial consistency during field-spectrum mapping.

[0075] This embodiment avoids signal crosstalk between microwave channels through a strict timing control strategy and clear frequency band definition, ensuring the purity of scattering data and the solvability of the physical model. At the same time, by utilizing the high-frequency acquisition characteristics of laser-induced breakdown spectroscopy, it is possible to capture the instantaneous compositional changes on the surface of rapidly moving coal flow, providing high temporal resolution boundary data for subsequent field-spectrum mapping.

[0076] Example 3:

[0077] Based on multi-angle microwave scattering parameters, a nonlinear physical field inversion operation is performed to reconstruct the complex permittivity distribution map of the cross-section of the coal flow to be measured in the furnace. This includes: calibrating and denoising the multi-angle microwave scattering parameters to generate standard scattering data; inputting the standard scattering data into a preset nonlinear tomographic imaging algorithm model, and minimizing the residual between the measured and calculated values ​​through iterative calculation; when the residual converges to a preset threshold range, outputting the complex permittivity distribution map of the current iteration step, where the real part of the complex permittivity represents the coal flow density and the imaginary part represents the local moisture content.

[0078] This embodiment further defines the specific algorithm flow for the physical field inversion operation; performs data preprocessing to calibrate the acquired raw microwave scattering parameters; normalizes the reference data under no-load conditions and uses wavelet transform to remove high-frequency electromagnetic noise to generate standard scattering data. Perform iterative inversion by inputting standard scattering data into a preset nonlinear tomographic imaging algorithm model. In this embodiment, the Born iterative method is used.

[0079] The core of this algorithm lies in solving the nonlinear integral equation and iteratively updating the distribution of the complex permittivity to be solved. To ensure the executability of the algorithm, the initial conditions of the iterative process are... The dielectric constant of the background medium is set as follows: Specifically, in each iteration step n, the system calculates the scattered field based on the current dielectric constant distribution estimate using a forward solver based on the discretized Lippmann-Schwinger equations. The calculation formula is as follows:

[0080]

[0081] in, The Green's function is given by the background medium. For the step of numerically solving using the method of moments, this embodiment constructs the following specific discrete linear equation system to analyze the total field vector within the imaging region. , dimension The calculation formula is as follows:

[0082]

[0083] in, It is the identity matrix. The scattering interaction matrix has its off-diagonal elements. The calculation formula is:

[0084]

[0085] diagonal elements The equivalent calculation using a cylinder is as follows:

[0086]

[0087] in, The equivalent radius of a square pixel unit after it has been converted into a circular region is used to approximate the self-actualization term; For pixel area, For a Bessel function of the first kind, The function is a Hankel function of the second kind; the accurate internal total field is obtained by solving the above dense linear equations using the LU decomposition method or the BiCGSTAB iterative method. This is used to calculate the new Jacobian matrix in the next iteration, thereby gradually approximating the non-uniform medium distribution;

[0088] These represent the index numbers of discrete pixels within the imaging area; This represents the index number of the receiving antenna. This represents the total number of receiving antennas. The imaginary unit, For the background wavenumber, is the Euclidean distance between pixels; For the main field, The total number of discrete pixels; in the formula This represents the spatial location vector of the field point under investigation. Indicates the first The center position vector of each discrete pixel. This indicates the position of the transmitting antenna when no scatterer is present. The incident electric field generated at the location;

[0089] Regarding the physical meaning of the formula parameters and the numerical discretization process, further clarification is provided here to eliminate ambiguity: Regarding It is defined as a pixel area micro-element after discretization of a two-dimensional imaging slice; this is because in the process of establishing the three-dimensional physical distribution model, this scheme prioritizes the use of tomography to invert the cross-section, i.e., two-dimensional, and stacks it axially along the coal conveying direction. Therefore, using two-dimensional area micro-elements in the kernel of the inversion algorithm is logical and accurate; regarding vectors The construction requires clarifying the numerical mapping process from continuous field functions to discrete vectors: dividing the imaging region into... After one grid, the continuous incident field function Discretely sampled as column vectors, its calculation formula is:

[0090]

[0091] in, For the first The center coordinates of each pixel are determined, ensuring a strict numerical match between the aforementioned linear equations and preventing confusion during algorithm replication. The Fréser derivative of the scattered field with respect to the objective function is calculated, and the Jacobian matrix is ​​constructed. Its elements Characterizing the first The dielectric constant perturbation of the nth pixel affects the first pixel. The effect of scattered field at each receiving antenna is calculated using the following formula:

[0092]

[0093] By solving the Tikhonov-regularized linear equation system Obtain the update value of the objective function. , Let be the update vector of the complex permittivity distribution to be solved, i.e. Its dimensions and total number of pixels Consistent; Updated dielectric constant distribution ; Calculate the updated residuals The calculation formula is as follows:

[0094]

[0095] in, Denotes the Euclidean norm. It is the conjugate transpose matrix. The regularization parameter is used; to ensure the convergence stability of the inversion process, this embodiment uses the Levenberg-Marquardt strategy to dynamically adjust the regularization parameter. In each iteration step, if the current residual If it increases compared to the previous step, then it will increase. Values, for example, multiplied by a factor of 10, to strengthen constraints; if residuals Decrease, then decrease The value can be adjusted by dividing by a factor of 2 to accelerate convergence; the above adjustment factor is an empirical setting, and in practical applications it can be adjusted according to the inversion convergence speed. Dynamically selected within the interval; initial The value is set to 0.01 times the largest eigenvalue of the Jacobian matrix;

[0096] The standard scattering data is derived from the preprocessing step, and its physical meaning is the actual measured electromagnetic response;

[0097] Derived from calculations using the forward electromagnetic scattering operator, its physical meaning is the scattering parameters estimated based on the current model;

[0098] It originates from interpolation operations and its physical meaning is the iterative convergence criterion.

[0099] Response to residual Once the convergence reaches a preset threshold range, the system stops iterating and outputs the complex permittivity distribution map for the current step; in this distribution map, the real part of the complex permittivity is... Primarily determined by the density of coal, it is used to characterize the density distribution of coal flow; the imaginary part of the complex permittivity. It is mainly determined by the loss of the medium and is used to characterize the local moisture content;

[0100] To achieve precise quantification of physical quantities, this embodiment pre-defines a dielectric-physical parameter conversion model calibrated through experiments; specifically, it establishes a linear mapping relationship between the real part of the complex permittivity and the density:

[0101]

[0102] And the polynomial mapping relationship between the imaginary and real parts of the complex permittivity and the moisture content:

[0103]

[0104] The above coefficients The model parameters were obtained by microwave regression testing of standard coal samples with varying density and moisture content for typical coal types, thus ensuring the physical authenticity of the model parameters.

[0105] This embodiment uses a nonlinear iterative algorithm to overcome the multiple scattering effect of microwaves in non-uniform media, and can reconstruct the complex structure inside the coal flow with high precision. By clearly distinguishing the physical meaning of the real part and the imaginary part of the complex permittivity, the system can simultaneously obtain the density field and the moisture field, realizing the decoupled characterization of multidimensional coal quality characteristics by a single physical field.

[0106] Example 4:

[0107] Using surface element characteristic spectra as boundary constraints, combined with complex permittivity distribution maps, the chemical element distribution inside the coal flow to be tested is deduced through a field-spectrum mapping model. This includes: extracting local moisture values ​​from the surface region from the complex permittivity distribution map; using the local moisture values ​​to perform matrix effect compensation on the surface element characteristic spectra at the same spatial location, eliminating the nonlinear interference of moisture fluctuations on spectral intensity, and generating corrected element concentration data.

[0108] Using the corrected elemental concentration data as boundary values, a semi-supervised manifold learning algorithm is used to map the surface chemical information to the internal volume region of the coal flow to be tested along the density gradient direction of the three-dimensional physical distribution model. The matrix effect compensation is performed on the surface elemental feature spectrum of the same spatial location using local moisture values, including: constructing a moisture-spectral intensity attenuation database and fitting the spectral correction coefficient curves under different moisture contents.

[0109] Based on the local moisture value, the corresponding target correction coefficient is found in the spectral correction coefficient curve; the original intensity value of the surface element characteristic spectrum is multiplied by the target correction coefficient to obtain the corrected spectrum reflecting the true element concentration.

[0110] This embodiment further defines the specific implementation mechanism of the field-spectrum mapping model and matrix effect compensation, which is key to improving the accuracy of wet coal detection. In executing the matrix effect compensation step, the system pre-constructs a moisture-spectral intensity attenuation database and fits it to obtain the spectral intensity attenuation ratio curve. ;

[0111] This curve characterizes the quenching effect of moisture on the plasma signal. Its mathematical form is set as a negative exponential decay model, and its calculation formula is as follows:

[0112]

[0113] After normalization, the product is made to meet the drying baseline conditions. , The fitting constant is determined as follows: Typical coal samples from the target mining area are selected and prepared into five groups of standard samples with different moisture gradients. The intensity of the characteristic spectral lines of each sample is measured under constant laser energy. Normalization is performed using the intensity of the dried sample as a benchmark to obtain the fitting constant. The observed values ​​were solved using nonlinear regression with the least squares method.

[0114] As a typical example, for bituminous coal, the range of constant values ​​obtained from the fitting is usually as follows: The value is typically 0.8-0.9, with a typical value of 0.85; attenuation coefficient. The unit is The value is taken as 0.10-0.15, with a typical value of 0.12; residual constant. Take a value of 0.1-0.2, with a typical value of 0.15, and satisfy the following conditions: Normalization constraints; extracting local moisture values ​​at corresponding spectral detection points from the complex permittivity distribution map. Based on the local moisture value, calculate the corresponding signal residual rate in the curve and obtain the target correction coefficient. That is, the reciprocal of the residual rate, is corrected as follows:

[0115]

[0116]

[0117] in, Derived from calibration calculations, its physical meaning is the calibrated elemental spectral intensity restored to the dry reference. Originating from a spectral detector, its physical meaning is the original measured intensity; Derived from reciprocal operations, its physical meaning is the spectral compensation gain coefficient; specifically, in order to obtain the spectral intensity... After obtaining the elemental concentration data from the example, the system performs an intensity-concentration calibration conversion: calling a preset calibration model. The corrected light intensity is converted into the dry basis mass percentage concentration of the element, such as the dry basis carbon content. %

[0118] Since the aforementioned matrix effect compensation step has removed the suppressive effect of moisture on the spectrum, the concentration obtained at this point characterizes the essential properties of the coal skeleton. This calibration model is obtained by partial least squares regression training on gradient concentration standard coal samples, whose dry basis composition has been pre-determined. Specifically, it is in linear polynomial form, and its calculation formula is as follows:

[0119]

[0120] in, Characteristic spectral lines The correction strength; as a specific example, for carbon content, the formula is specified as:

[0121]

[0122] in, and The carbon atom emission line correction intensities at 247.8 nm and 193.0 nm, respectively, ensured the subsequent boundary values. It has a clearly defined dry basis physical dimension unit (%), thus supporting subsequent calorific value calculations; based on this, a semi-supervised manifold learning mapping step is performed to obtain the corrected surface element concentration. Then, a density manifold space is established using a semi-supervised manifold learning algorithm;

[0123] This step is based on the assumption that physical properties are similar in the same type of coal and that spatially adjacent regions have chemical similarities. It propagates the chemical information of known points on the surface to unknown points in the interior along the density gradient direction of the three-dimensional physical distribution model.

[0124] Specifically, the semi-supervised manifold learning algorithm employs the density-weighted graph Laplacian regularization method; the system constructs a graph structure with coal flow micro-elements as nodes, defining topological constraints for spatial adjacency: for any micro-element node in the 3D mesh Only the infinitesimal elements within its 26-neighborhood in space are defined as adjacent nodes. The connection weights of the remaining non-contact nodes are forcibly set to zero to ensure that the propagation of chemical information strictly follows the continuity of physical entities;

[0125] Define adjacent infinitesimal elements Weights between Considering the potential physical ambiguity of a single density index in heterogeneous coal mixtures—for example, high-ash coal and high-moisture coal may have similar densities, causing the homogeneity assumption to fail—this embodiment introduces moisture distribution obtained synchronously through microwave inversion to ensure the uniqueness and accuracy of chemical property extrapolation. As a second feature constraint, a density-moisture joint feature space is constructed; the modified weighting function is defined as:

[0126]

[0127] in, This refers to the one-dimensional index number of the micro-element in the three-dimensional mesh model; These are the diffusion coefficients for density and moisture, respectively. Physically, they are used to control the smoothness of information propagation in the manifold space, i.e., the attenuation rate. Their values ​​are set to 0.5 to 1.0 times the standard deviation of the density distribution and the standard deviation of the moisture distribution of the current batch of coal.

[0128] The theoretical basis for this joint characteristic constraint lies in the essential difference between the dielectric response mechanisms of coal and rock: as an inorganic mineral, gangue has a high density but its dielectric loss factor, i.e., the imaginary part, is extremely low in the microwave frequency band, exhibiting the characteristics of a high real part and a low imaginary part.

[0129] While high-moisture coal lumps may have a density similar to gangue, exhibiting a similar solid part, their dielectric loss factor is significantly higher than that of dry rocks due to the polar relaxation effect of water molecules, thus exhibiting characteristics of high solid part and high imaginary part.

[0130] By introducing a moisture distribution, derived primarily by the imaginary part, as a second-dimensional constraint orthogonal to density, the aforementioned manifold learning algorithm can mathematically and effectively separate wet coal from dry rock in the feature space. This avoids erroneous mapping of chemical information caused by relying solely on density gradients, providing solid rock physics support for the field-spectrum mapping model. The energy function is calculated by minimizing the energy function, as follows:

[0131]

[0132] in, To determine the concentration of the element to be determined, the chemical composition tends to propagate along the joint characteristic isosurface. To obtain the globally optimal solution for this energy function, this embodiment constructs a graph Laplacian matrix, the calculation formula of which is:

[0133]

[0134] in, For a degree matrix, the diagonal elements The nodes are divided into a set of labeled surface nodes. Its value is taken from the above. and unlabeled internal node set Therefore, the energy minimization problem is mathematically transformed into solving a system of linear equations. ; Matrix inversion operation The element concentration vector of the unknown internal region can be directly calculated analytically. This mathematically achieves a precise mapping along the density gradient direction;

[0135] This embodiment cleverly utilizes the precise moisture value obtained from microwave inversion to correct spectral errors, turning a disadvantage into an advantage. Through reciprocal compensation and PLS calibration algorithms, it effectively solves the problems of nonlinear signal decline and dimensionless loss caused by moisture quenching plasma in wet coal detection using laser-induced breakdown spectroscopy. At the same time, by using density field and moisture field as joint guides, surface chemical information is scientifically introduced into the interior. Compared with the simple surface averaging method, it is more in line with the statistical laws of coal geology and significantly improves the confidence of internal elemental extrapolation.

[0136] Example 5:

[0137] Based on the established three-dimensional physical distribution model and the deduced chemical element distribution, a volume-weighted integral algorithm is used for coal quality analysis, including: dividing the coal flow to be tested into multiple micro-elements; determining the density weight of each micro-element based on the complex dielectric constant distribution diagram; determining the elemental component concentration of each micro-element based on the chemical element distribution; and performing density-weighted volume integral on the elemental component concentrations of all micro-elements to calculate the weighted average calorific value, total sulfur content, and total ash content of the entire coal flow to be tested.

[0138] This embodiment further defines the specific calculation process for coal quality analysis; the coal flow to be tested is logically divided into... There are three infinitesimal volume elements, each with a volume of . For the first Each infinitesimal element has its density weight obtained from the complex permittivity distribution diagram. and moisture content The concentration of its dry basis elemental components is obtained from the chemical element distribution and denoted as . This data originates from the field-spectrum mapping model's derivation of the dry basis boundary values; at this point, the system performs a benchmark conversion step, utilizing the local moisture value of this micro-element provided by the complex permittivity distribution map. Calculate the concentration of the received basic elements: ,in Representing each element;

[0139] Regarding the required oxygen content in the formula The system performs density-spectral coupled regression calculations on the final output total ash content index. Given the significant positive correlation between coal density and ash content (i.e., mineral density is greater than organic matter density) and the negative correlation with carbon content, the system calls a pre-defined density-carbon-ash ternary regression model to calculate the dry-basis ash content of this micro-element. The calculation formula is as follows:

[0140]

[0141] in, The regression coefficients were determined for the specific coal type in this mining area. To ensure that the calculation results conform to physical principles, i.e., the ash content is non-negative and within a reasonable range, the system used large sample data to rigorously verify the dimensions and perform numerical fitting on the coefficients. Through statistical regression of no less than 50 sets of samples, the specific values ​​of the above coefficients were determined as follows: The unit is This significantly demonstrates the strong positive correlation between density and ash content; , is a dimensionless coefficient, reflecting the negative correlation fine-tuning of carbon content; The unit is , as the corrected intercept;

[0142] For example, when density That is, high-ash gangue with a carbon content At that time, the ash content was calculated to be approximately When density carbon content At that time, the ash content was calculated to be approximately This set of coefficients ensures the physical validity of the calculation results across the entire domain; the oxygen content on a dry basis is estimated using the difference method, and the calculation formula is as follows:

[0143]

[0144] Given nitrogen The quantity is low and fluctuates little. In this embodiment, a clear priority strategy is adopted to determine its value: the system first reads the set value in the initialization configuration file, which is 1.0% by default. If it is not set or the reading fails, the historical average value of 1.08% from the annual geological analysis report of the target mining area is automatically called, thereby eliminating the ambiguity of the value selection and further converting it to obtain the basic oxygen content. The comprehensive parameters of the entire coal flow are calculated using a volume-weighted integral algorithm; the weighted average received lower heating value is used as the basis. For example, the calculation formula is as follows:

[0145]

[0146] in, Derived from integral calculation, its physical meaning is the weighted average net calorific value of the entire coal flow.

[0147] Derived from elemental composition calculations, its physical meaning is the first The calorific value per unit mass of a microelement under received baseline conditions; its specific calculation is based on the elemental concentration within the microelement, and is solved using the modified Dulong formula, the calculation formula being:

[0148]

[0149] The coefficients are calibrated based on the specific coal type to ensure that the calculation results are lower heating values. The calibration method of the coefficients is as follows: collect no less than 30 sets of calibration coal samples covering the coal quality fluctuation range of the mining area, and use standard laboratory methods to determine their actual lower heating values ​​and the content of each element; construct a multiple linear regression equation, with the laboratory measured calorific value as the target variable and the element and moisture content as independent variables, and correct the above empirical coefficients through regression analysis.

[0150] In the absence of specific calibration data, the coefficients in the above formula adopt the generally recommended values ​​of Mendeleev's empirical formula, where the coefficient 339 corresponds to the heat contribution of carbon and 1030 corresponds to the heat contribution of available hydrogen. The specific values ​​of the coefficients can be fine-tuned according to the volatile matter content of the coal.

[0151] Derived from the product of density and volume, its physical meaning is the first... The actual mass of each micro-element is the received basic mass; similarly, the total sulfur content and total ash content are calculated using a similar mass-weighted integral logic.

[0152] This embodiment strictly adheres to the principle of consistency of physical dimensions, abandoning the coarse processing that assumes uniform coal flow density in traditional detection methods. By first calculating the specific energy of each micro-element and then performing global integration based on actual mass weights, it accurately considers the weighted impact of changes in porosity and bulk density within the coal flow on the contribution of calorific value. In scenarios where coal quality fluctuates significantly, this integration strategy can significantly reduce the weighted calculation error caused by uneven density distribution, thereby improving the commercial settlement accuracy of the final coal quality settlement data.

[0153] Example 6:

[0154] The method also includes: real-time monitoring of dielectric constant abrupt change regions in the complex dielectric constant distribution map; wherein, the preset metal threshold is greater than the preset rock threshold; if the dielectric constant value of the abrupt change region is greater than the metal threshold, it is determined that there are metal foreign objects in the coal flow to be tested, and an rejection indication signal is generated.

[0155] If the dielectric constant of the abrupt change region is less than or equal to the metal threshold and greater than the rock threshold, it is determined that there are large pieces of gangue in the coal flow to be tested, and a warning signal is generated; if the dielectric constant of the abrupt change region is less than or equal to the rock threshold, it is determined to be normal coal quality, and the coal quality analysis steps continue.

[0156] This embodiment further adds a foreign object identification and graded early warning mechanism based on dielectric constant abrupt changes; the system presets two key thresholds: a metal threshold. Set to an extremely high value, corresponding to the total internal reflection characteristics of a good metallic conductor; rock threshold. Set to medium to high values, corresponding to the dielectric properties of gangue or large rocks;

[0157] The aforementioned thresholds were not arbitrarily set, but obtained through statistical calibration; among them, the rock threshold... Set as the statistical distribution of dielectric constant of normal coal flow Upper limit value, to distinguish between coal quality fluctuations and rock abrupt changes;

[0158] Metal threshold Set as the lower limit of the equivalent dielectric constant for the reflective properties of pure metals in the microwave band, for example, in At operating frequency, set and It should be noted that this threshold is related to the operating frequency. If the system operating frequency changes, this physical threshold needs to be recalibrated.

[0159] The concept of equivalent dielectric constant is introduced here because in the finite element medium inversion algorithm, the strong conductive reflection characteristics of metals are numerically manifested as an abnormally high saturation of the real part of the dielectric constant. This does not contradict the physical definition of the complex dielectric constant, but is a characteristic representation at the algorithm level.

[0160] This value is significantly higher than the upper limit of the real part of the dielectric constant of high-moisture coal, and is usually less than 40, thus having clear distinguishability. By utilizing the high reflectivity and high loss characteristics of metals, it is ensured that there is a significant distinction from the dielectric properties of high-moisture coal.

[0161] The system scans each infinitesimal point in the complex permittivity distribution map in real time and extracts the real part of its complex permittivity. Density, as a characteristic quantity, is used as a criterion to identify regions of abrupt changes in dielectric constant; the real part of the response to the abrupt change region... The system determines that there are metal foreign objects such as detonators and iron parts in the coal flow to be tested, and immediately generates a rejection indication signal to trigger the action of the iron separator.

[0162] In response to The system detects the presence of large pieces of gangue, generates a warning signal, and alerts the coal mill to load fluctuations; in response to The system determined that the coal was of normal quality and continued with subsequent coal quality analysis steps.

[0163] This embodiment fully utilizes the difference in dielectric sensitivity of microwaves to metals and rocks, and clearly defines the real part of the complex dielectric constant as the criterion. While performing coal quality testing, it also embeds a foreign object detection function. This multi-task parallel processing mechanism not only effectively protects the downstream coal mill equipment and avoids equipment damage caused by metal foreign objects, but also eliminates the interference of foreign objects on calorific value calculation, ensuring the purity of the coal quality data entering the furnace.

[0164] Example 7:

[0165] After outputting the comprehensive coal quality parameters, the process also includes: comparing the comprehensive coal quality parameters with the preset combustion control model; if the weighted average calorific value in the comprehensive coal quality parameters is lower than the preset combustion efficiency threshold, generating an adjustment feedback signal for the coal mill speed or coal feed rate; and transmitting the adjustment feedback signal to the boiler distributed control system to achieve closed-loop combustion control.

[0166] This embodiment further defines the application logic of the detection results in the combustion closed-loop control; after outputting the comprehensive coal quality parameters, the system performs a comparison step, comparing the real-time calculated weighted average calorific value with the preset combustion control model in the DCS system; a feedback generation step is performed, in response to the weighted average calorific value in the comprehensive coal quality parameters being lower than the preset combustion efficiency threshold, indicating a deterioration in the quality of the coal entering the furnace; the system uses a discrete PID control algorithm to generate an adjustment feedback signal: setting the target calorific value benchmark as... The real-time deviation is calculated using the following formula:

[0167]

[0168] The formula for calculating the adjustment amount of the coal feeder speed is as follows:

[0169]

[0170] in, The proportional, integral, and derivative gain coefficients are preset. To clarify the source of these key control parameters and ensure the feasibility of the solution, this embodiment uses the critical proportional gain method, i.e., the Ziegler-Nichols closed-loop tuning method, to determine the specific values. During the system debugging phase, the integral... and differential Set to 0, then gradually increase the proportional gain. Until the system output, i.e., the thermal response, produces constant-amplitude oscillations, the critical gain at this point is recorded. and oscillation period ;

[0171] Calculated according to PID tuning rules: , , As a verified specific example, for a typical 600MW coal-fired power generating unit, the measured values ​​were... Thus setting ;like The command increases the coal feed rate or adjusts the coal mill speed to maintain the boiler load; the signal transmission step is executed, and the adjustment feedback signal is transmitted to the boiler distributed control system via industrial Ethernet to achieve closed-loop combustion control;

[0172] This embodiment elevates the application of detection data from traditional post-event settlement to the real-time control level; through second-level detection feedback, the boiler control system can perform feedforward control for coal quality fluctuations. This rapid response mechanism effectively mitigates combustion instability caused by sudden changes in coal quality, improves boiler combustion efficiency, and reduces the risk of unplanned shutdowns.

[0173] Example 8:

[0174] Please see Figure 2 An online coal quality detection system for furnace feed includes: a data acquisition unit, used to acquire multi-angle microwave scattering parameters and surface element characteristic spectrum of the coal flow to be tested through a microwave tomography sensor array and a surface spectral detector; and a physical field reconstruction unit, used to perform nonlinear physical field inversion calculation based on the multi-angle microwave scattering parameters, reconstruct the complex permittivity distribution map of the cross section of the coal flow to be tested, and establish a three-dimensional physical distribution model including density distribution and moisture distribution.

[0175] The field-spectrum coupling analysis unit is used to use the surface element characteristic spectrum as a boundary constraint condition, combined with the complex permittivity distribution map, to deduce the chemical element distribution inside the coal flow to be tested through the field-spectrum mapping model; the comprehensive parameter solution unit is used to perform coal quality analysis based on the established three-dimensional physical distribution model and the deduced chemical element distribution, and output comprehensive coal quality parameters.

[0176] This embodiment details the hardware device architecture for implementing the above method. The system mainly includes a core logic unit running in an industrial computer or embedded processor; the data acquisition unit is physically connected to the microwave sensor array and the laser-induced breakdown spectroscopy probe, and is responsible for the underlying drive control and synchronous acquisition of raw data.

[0177] The physical field reconstruction unit has a built-in electromagnetic field inverse scattering algorithm library, which is used to perform nonlinear physical field inversion operations and output complex permittivity distribution map; the field spectrum coupling analysis unit is the core processing module, which is used to perform matrix effect compensation and manifold learning mapping, and integrate physical and chemical information; the comprehensive parameter solution unit performs volume integration operations, outputs the final industrial analysis indicators, and has a foreign object alarm interface.

[0178] This embodiment adopts a modular system architecture design, which decouples the various functional units and facilitates algorithm upgrades and maintenance. The close cooperation between hardware and software ensures the efficient implementation of the detection method. Especially in complex industrial environments, this specialized unit design ensures the stability and real-time performance of the system when processing high-throughput data.

[0179] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for online detection of coal quality entering the furnace, characterized in that, Includes the following steps: The microwave tomography sensor array and surface spectral detector were used to obtain the multi-angle microwave scattering parameters and surface element characteristic spectra of the coal flow to be tested. Based on the multi-angle microwave scattering parameters, a nonlinear physical field inversion operation is performed to reconstruct the complex permittivity distribution map of the cross-section of the coal flow to be tested, and a three-dimensional physical distribution model including density distribution and moisture distribution is established. Using the surface elemental characteristic spectrum as a boundary constraint, and combining it with the complex permittivity distribution map, the chemical element distribution inside the coal stream to be tested is deduced through a field-spectrum mapping model, including: Extract the local moisture value of the surface region from the complex permittivity distribution map; The local moisture value is used to perform matrix effect compensation on the surface element characteristic spectrum at the same spatial location, eliminating the nonlinear interference of moisture fluctuation on spectral intensity and generating corrected element concentration data. Using the corrected elemental concentration data as boundary values, a semi-supervised manifold learning algorithm is used to map the surface chemical information to the internal volume region of the coal flow to be tested along the density gradient direction of the three-dimensional physical distribution model. Based on the established three-dimensional physical distribution model and the derived chemical element distribution, a volume-weighted integral algorithm is used to perform coal quality analysis and output comprehensive coal quality parameters.

2. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, The process involves acquiring multi-angle microwave scattering parameters and surface elemental characteristic spectra of the coal flow to be tested using a microwave tomography sensor array and a surface spectral detector, including: The microwave tomography sensor array surrounding the coal conveying channel is controlled to transmit and receive microwave signals that penetrate the coal flow to be tested in sequence according to a preset time sequence, and the scattering parameter matrix under different projection angles is recorded. A laser-induced breakdown spectroscopy probe positioned above the coal flow to be tested is controlled to acquire the plasma emission spectrum of the surface layer of the coal flow at high frequency, generating a surface elemental characteristic spectrum containing carbon, hydrogen, and sulfur characteristic peaks.

3. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, The process of performing nonlinear physical field inversion calculations based on the multi-angle microwave scattering parameters to reconstruct the complex permittivity distribution map of the cross-section of the coal flow entering the furnace includes: The multi-angle microwave scattering parameters are calibrated and denoised to generate standard scattering data; The standard scattering data is input into a preset nonlinear tomographic imaging algorithm model, and the residual between the measured value and the calculated value is minimized through iterative calculation. When the residual converges to a preset threshold range, the complex permittivity distribution map of the current iteration step is output, where the real part of the complex permittivity represents the coal flow density and the imaginary part represents the local moisture content.

4. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, The aforementioned analysis of coal quality using a volume-weighted integral algorithm, based on the established three-dimensional physical distribution model and the derived chemical element distribution, includes: The coal flow to be tested is divided into multiple micro-elements; The density weight of each of the micro-elements is determined based on the complex permittivity distribution diagram. The elemental composition concentration of each micro-element is determined based on the chemical element distribution. Density-weighted volume integrals are performed on the elemental component concentrations of all the aforementioned micro-elements to calculate the weighted average calorific value, total sulfur content, and total ash content of the entire coal flow to be tested.

5. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, Also includes: Real-time monitoring of abrupt changes in dielectric constant in the complex dielectric constant distribution diagram; Among them, the preset metal threshold is greater than the preset rock threshold; If the dielectric constant of the abrupt change region is greater than the metal threshold, it is determined that there is a metal foreign object in the coal flow to be tested, and an rejection indication signal is generated. If the dielectric constant of the abrupt change region is less than or equal to the metal threshold and greater than the rock threshold, it is determined that there are large pieces of gangue in the coal flow to be tested, and a warning signal is generated. If the dielectric constant of the abrupt change region is less than or equal to the rock threshold, it is determined to be normal coal quality, and the coal quality analysis steps continue.

6. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, Following the output of the comprehensive coal quality parameters, the following is also included: The comprehensive coal quality parameters are compared with the preset combustion control model; If the weighted average calorific value in the comprehensive coal quality parameters is lower than the preset combustion efficiency threshold, an adjustment feedback signal for the coal mill speed or coal feed rate is generated. The adjustment feedback signal is transmitted to the boiler distributed control system to achieve closed-loop combustion control.

7. The method for online detection of coal quality entering the furnace according to claim 1, characterized in that, The method of using the local moisture value to perform matrix effect compensation on the surface elemental characteristic spectrum at the same spatial location includes: A moisture-spectral intensity attenuation database was constructed, and spectral correction coefficient curves under different moisture contents were obtained by fitting. Based on the local moisture value, find the corresponding target correction coefficient in the spectral correction coefficient curve; Multiply the original intensity value of the surface element characteristic spectrum by the target correction coefficient to obtain a corrected spectrum that reflects the true element concentration.

8. An online coal quality detection system for furnace feed, employing the online coal quality detection method as described in any one of claims 1-7, characterized in that, include: The data acquisition unit is used to acquire the multi-angle microwave scattering parameters and surface element characteristic spectrum of the coal flow to be tested through a microwave tomography sensor array and a surface spectral detector, respectively. The physical field reconstruction unit is used to perform nonlinear physical field inversion calculations based on the multi-angle microwave scattering parameters, reconstruct the complex permittivity distribution map of the cross section of the coal flow to be tested, and establish a three-dimensional physical distribution model including density distribution and moisture distribution. The field-spectrum coupling analysis unit is used to use the surface element characteristic spectrum as a boundary constraint condition, and combined with the complex permittivity distribution map, to deduce the chemical element distribution inside the coal flow to be tested through the field-spectrum mapping model; The comprehensive parameter calculation unit is used to perform coal quality analysis using a volume-weighted integral algorithm based on the established three-dimensional physical distribution model and the derived chemical element distribution, and output comprehensive coal quality parameters.

Images