Method and system for analyzing coal dust mass data

Abstract

The application discloses a coal powder quality data analysis method and system, and belongs to the technical field of data analysis. Through systematic collaborative processing of multi-source heterogeneous data, the problem of insufficient evaluation dimensions caused by single data source in related coal powder quality analysis methods is solved, the cross-validation mechanism of material property correction parameters and particle characterization features improves the physical accuracy of element concentration correction and avoids the interference of matrix effect; the dynamic generation of working condition adaptation parameters makes the data processing intensity and the actual running state match in real time, and enhances the working condition adaptability of the analysis result; the weight dynamic allocation mechanism driven by the distribution stability parameter ensures that the decision fusion process can adapt to the change of the coal powder flow state, and overcomes the defect of weak real-time matching capability; the multi-objective optimization model takes quality stability and energy consumption minimization as cooperative targets, and the generated pulverizing system optimization parameters meet the quality requirements and energy efficiency optimization at the same time, thereby solving the problem that optimization guidance cannot be generated considering quality stability and system energy consumption.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to a method and system for analyzing coal powder quality data. Background Technology

[0002] In industrial sectors such as coal-fired power generation and coal chemical industry, the quality and conveying status of pulverized coal directly affect combustion efficiency, equipment safety, and pollutant emissions. Related technologies typically employ single or a few detection methods for offline analysis or partial online monitoring of pulverized coal, such as using spectral analysis of elemental composition or determining flow state through charge signals.

[0003] The existing methods cannot provide a comprehensive, accurate, and real-time assessment of pulverized coal quality that matches operating conditions, and it is even more difficult to generate optimized operating guidelines that can simultaneously consider quality stability and system energy consumption. This results in the pulverizing system often operating in a suboptimal state, hindering the improvement of overall energy efficiency and the control of stable production.

[0004] Therefore, there is an urgent need in this field for a coal powder quality analysis method that can deeply integrate multi-source heterogeneous data and fully adapt to changes in operating conditions, in order to solve the problems of single analysis dimensions, insufficient data utilization, and one-sided decision-making in related technologies. Summary of the Invention

[0005] This application provides a method and system for analyzing pulverized coal quality data, which can solve the problems of single analysis dimension, insufficient data utilization, and biased decision-making. The technical solution is as follows: On the one hand, a method for analyzing pulverized coal quality data is provided, the method comprising: In response to the command to analyze the quality of coal powder in the target pipeline, multi-source heterogeneous data of coal powder is acquired. The multi-source heterogeneous data of coal powder includes atomic emission spectroscopy data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data, and dynamic parameters of the pulverizing process. The atomic emission spectrum data and the three-dimensional holographic reconstruction data are subjected to the first type of data processing to obtain material property correction parameters and particle characterization features, and operating condition adaptation parameters are generated based on the dynamic parameters of the powder making process. The coal flow acoustic data, the space charge distribution data, and the particle characterization features are subjected to a second type of data processing to obtain coal powder flow uniformity index and distribution stability parameters; Based on the material property correction parameters, the operating condition adaptation parameters, the flow uniformity index, and the distribution stability parameters, coal powder quality index analysis results and pulverizing system optimization parameters are generated through multi-source decision fusion. The weight allocation of each data source in the multi-source decision fusion is dynamically adjusted based on the distribution stability parameters.

[0006] On the one hand, a coal powder quality data analysis system is provided, the system comprising: The acquisition module is used to acquire multi-source heterogeneous data of coal powder in response to the quality analysis command of coal powder in the target pipeline. The multi-source heterogeneous data of coal powder includes atomic emission spectrum data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data and dynamic parameters of the pulverizing process. The first processing module is used to perform a first type of data processing on the atomic emission spectrum data and the three-dimensional holographic reconstruction data to obtain material property correction parameters and particle characterization features, and to generate working condition adaptation parameters based on the dynamic parameters of the powder making process. The second processing module is used to perform a second type of data processing on the coal flow acoustic data, the space charge distribution data, and the particle characterization features to obtain coal powder flow uniformity index and distribution stability parameters. The generation module is used to generate coal powder quality index analysis results and pulverizing system optimization parameters based on the material property correction parameters, the operating condition adaptation parameters, the flow uniformity index, and the distribution stability parameters through multi-source decision fusion. The weight allocation of each data source in the multi-source decision fusion is dynamically adjusted based on the distribution stability parameters.

[0007] On one hand, a computer device is provided, the computer device including one or more processors and one or more memories, the one or more memories storing at least one computer program, the computer program being loaded and executed by the one or more processors to implement the method for analyzing the coal powder quality data.

[0008] On one hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, the computer program being loaded and executed by a processor to implement the method for analyzing the coal powder quality data.

[0009] On one hand, a computer program product or computer program is provided, which includes program code stored in a computer-readable storage medium. The processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to perform the aforementioned method for analyzing coal powder quality data. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the implementation environment of a method for analyzing coal powder quality data provided in an embodiment of this application; Figure 2 This is a flowchart of a method for analyzing pulverized coal quality data provided in an embodiment of this application; Figure 3 This is a partial flowchart of a method for analyzing pulverized coal quality data provided in an embodiment of this application; Figure 4 This is a partial flowchart of another method for analyzing pulverized coal quality data provided in an embodiment of this application; Figure 5 This is a partial flowchart of another method for analyzing pulverized coal quality data provided in an embodiment of this application; Figure 6 This is a partial flowchart of another method for analyzing pulverized coal quality data provided in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of a coal powder quality data analysis system provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of a server provided in an embodiment of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0013] In this application, the terms "first," "second," etc., are used to distinguish identical or similar items with essentially the same function. It should be understood that there is no logical or temporal dependency between "first," "second," and "nth," nor are there any restrictions on quantity or execution order.

[0014] First, some terms used in the embodiments of this application will be introduced.

[0015] Atomic emission spectroscopy data refers to the spectral signals containing characteristic spectral lines of different elements collected after a coal powder sample is excited (e.g., by laser or plasma) to atomize and emit light through transitions. This data is the primary basis for analyzing the types of elements (such as carbon, hydrogen, potassium, sodium, etc.) and their concentrations in coal powder.

[0016] Three-dimensional holographic reconstruction data refers to the original data obtained using holographic imaging technology that reflects the distribution and morphology of coal powder particles in three-dimensional space. By processing this data, the three-dimensional shape and size of each particle, as well as the spatial positional relationship of all particles, can be reconstructed.

[0017] Coal flow acoustic signature data refers to the sound waves or vibration signals generated by the collision and friction between particles and between particles and the pipe wall during the flow of pulverized coal in a pipeline. The energy distribution and pattern characteristics of this data in a specific frequency band (such as the low frequency band) are closely related to the flow state of the pulverized coal (such as flow velocity and uniformity).

[0018] Space charge distribution data refers to the space charge signal generated by frictional charging during the flow of pulverized coal, obtained by simultaneously measuring multiple electrostatic probe arrays arranged along the circumference of the pipeline. This data reflects the distribution of pulverized coal concentration and flow state across the pipeline cross-section.

[0019] Dynamic parameters of the pulverizing process: These refer to the real-time operating data generated by the pulverizing system (such as coal mill and pulverizing fan) during operation, mainly including coal mill current, inlet and outlet pressure difference, primary air volume, and air pressure. These parameters directly reflect the load status of the equipment and the operational stability of the system.

[0020] Material property correction parameters: These are coefficients obtained by cross-validating atomic emission spectrometry data with three-dimensional holographic topography data, used to compensate for and correct interferences such as matrix effects and light scattering effects. Their function is to improve the accuracy of material property analysis results, such as elemental concentration.

[0021] Operating condition adaptation parameters: These refer to adjustment factors calculated by the model based on the dynamic parameters of the pulverizing process, used to dynamically adjust the intensity of the first and second types of data processing. This enables the entire analysis method to adapt to different operating conditions (such as high load and low load), ensuring the robustness of the analysis results.

[0022] Coal powder flow uniformity index: A quantitative index calculated by integrating acoustic and electrostatic characteristics, used to characterize whether the concentration distribution of coal powder is uniform across the pipe cross-section. The higher the index value, the more uniform the coal powder distribution, which is beneficial to stable combustion.

[0023] Distribution stability parameter: A quantitative parameter integrating temporal and spatial dimensions, used to characterize the temporal fluctuations and spatial synchronicity of pulverized coal flow. It reflects the stability of the flow process and is an important basis for dynamically adjusting data fusion weights.

[0024] Pulverizing system optimization parameters: These refer to a set of specific operating instructions or set values ​​(such as suggested coal mill output, primary air volume, etc.) obtained by solving a multi-objective optimization model. The goal is to minimize unit energy consumption while ensuring stable coal powder quality.

[0025] Acoustic characteristic parameters: These are the characteristic quantities that accurately reflect the flow state obtained after extracting frequency band energy from the original coal flow acoustic data and using particle morphology parameters (such as sphericity and roughness) for propagation attenuation compensation and normalization.

[0026] Cross-correlation coefficient: A statistic used to measure the similarity between signals from electrostatic probes at different locations when analyzing space charge distribution data. Its peak value reflects signal correlation, and the time delay corresponding to the peak value reflects the time difference of coal powder flowing through different probes.

[0027] Electrostatic characteristic parameters: These are comprehensive characteristic quantities used to quantify the uniformity of space charge distribution and flow stability, obtained by weighted fusion of cross-correlation coefficients from multiple probes (e.g., by correcting for particle surface characteristics).

[0028] Particle sphericity: A parameter describing how close the shape of pulverized coal particles is to an ideal sphere. It is a numerical value, with a value of 1 for perfectly spherical particles, and the smaller the value is as the shape becomes more irregular. It directly affects the flowability of the particles and the sound / light scattering characteristics.

[0029] Propagation attenuation coefficient: A physical parameter that quantifies the degree of energy attenuation of sound waves propagating in a pulverized coal gas-solid two-phase flow medium. Its value is related to the particle concentration, shape, and surface characteristics, and is used for inverse compensation of acoustic measurement signals.

[0030] Dynamic time warping: an algorithm for aligning two signal sequences that exhibit nonlinear scaling or velocity differences along the time axis. This is used to address the potential timing asynchrony between acoustic and electrostatic signals due to their different physical principles.

[0031] Extreme delay values: The maximum and minimum delays found in the cross-correlation analysis of all probe combinations. It reflects the fluctuation range of the pulverized coal flow velocity.

[0032] Time delay dispersion: A statistic (such as standard deviation) used to measure the degree of dispersion of time delay values ​​calculated by all probe combinations. It reflects the magnitude of the difference in pulverized coal flow velocity in different areas of the pipeline.

[0033] Time delay stability coefficient: A parameter used to quantify the spatial stability of pulverized coal flow, calculated by combining time delay extrema and time delay dispersion. The smaller the coefficient, the worse the spatial stability.

[0034] Particle position update: In the multi-objective particle swarm optimization algorithm, each "particle" (representing a potential solution) adjusts its position in the next iteration (i.e., a new set of powdering system operating parameters) based on its own historical best position and the global best position of the entire population.

[0035] Velocity update: In particle swarm optimization, the change in a particle's position is defined as "velocity". Velocity update determines the direction and step size of the particle's search in the solution space, and it is influenced by the particle's current velocity, individual cognition, and social cognition.

[0036] Pareto optimal solution set: In multi-objective optimization problems, this refers to the set of optimal solutions that cannot further improve any one objective function without compromising at least one of the other. In this specific context, each solution in this set represents a feasible operational scheme that achieves a certain balance between "coal powder quality stability" and "system unit energy consumption".

[0037] In coal-fired power generation and coal chemical pulverizing systems, coal powder quality analysis has long been limited by the single data source and insufficient adaptability to operating conditions. Related technologies rely solely on single detection methods such as atomic emission spectroscopy or space charge distribution data for offline analysis or localized monitoring, resulting in fragmented assessment dimensions and an inability to simultaneously acquire multi-dimensional information such as elemental composition, particle morphology, and flow state. Furthermore, the analysis process lacks real-time response to dynamic parameters of the pulverizing process, causing a disconnect between quality assessment results and operating conditions, making it difficult to generate optimized operational guidance that balances quality stability and system energy consumption. Consequently, the pulverizing system operates continuously in a suboptimal state, further highlighting problems such as reduced combustion efficiency, increased equipment safety risks, and greater difficulty in controlling pollutant emissions.

[0038] For example, during the pulverized coal transport process in a target pipeline of a coal-fired power plant, when the pulverizer load fluctuates due to grid dispatch demands, the relevant methods rely solely on atomic emission spectroscopy data for elemental composition analysis, failing to integrate three-dimensional holographic reconstruction data to capture particle surface morphology features, nor combining coal flow acoustic data and space charge distribution data to assess the flow state. In this situation, the airflow and current changes in the dynamic parameters of the pulverizing process are not converted into operating condition adaptation parameters, resulting in particle characterization characteristics failing to accurately reflect the correlation between chemical activity and flow characteristics; the coal flow acoustic energy distribution is distorted due to particle shape interference, and the space charge cross-correlation coefficient is biased due to surface characteristics, ultimately causing the calculated results of flow uniformity indicators and distribution stability parameters to deviate from the actual physical state. Specifically, the quality analysis results cannot match the current operating conditions, and optimization guidance focuses only on a single dimension, leading to combustion instability and increased system energy consumption.

[0039] If the above problems are not addressed, the one-sidedness of pulverized coal quality assessment will continue to constrain the operational efficiency of the pulverizing system. The lack of real-time matching capability between quality indicator analysis results and changes in operating conditions prevents optimized parameters from synergistically balancing quality stability and energy consumption control targets; consequently, combustion efficiency continues to deteriorate, equipment safety risks accumulate, pollutant emission fluctuations intensify, and the control targets of overall energy efficiency improvement and stable production are difficult to achieve.

[0040] Figure 1 This is a schematic diagram illustrating the implementation environment of a method for analyzing pulverized coal quality data provided in this application embodiment. See also... Figure 1 This implementation environment may include node 110 and server 140.

[0041] Node 110 is connected to server 140 via a wireless or wired network. Node 110 has an application installed and running that supports the analysis of pulverized coal quality data.

[0042] Server 140 is a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms. Server 140 can provide background services for applications running on node 110.

[0043] Since the embodiments of this application involve a relatively complex data processing process, the computing power of node 110 is insufficient to meet the requirements under the current computing power background. Therefore, the relevant data will be sent to server 140, and server 140 will execute the technical solution provided in the embodiments of this application.

[0044] To address the aforementioned technical problems, this application proposes a method for analyzing pulverized coal quality data, see [link to relevant documentation]. Figure 2 Taking the server as the executing entity as an example, the following steps are included.

[0045] 201. In response to the quality analysis command of coal powder in the target pipeline, acquire multi-source heterogeneous data of coal powder, including atomic emission spectrum data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data and dynamic parameters of the pulverizing process; 202. Perform the first type of data processing on atomic emission spectroscopy data and three-dimensional holographic reconstruction data to obtain material property correction parameters and particle characterization features, and generate working condition adaptation parameters based on the dynamic parameters of the powder making process; 203. Perform second-class data processing on coal flow acoustic data, space charge distribution data and particle characterization characteristics to obtain coal powder flow uniformity index and distribution stability parameters; 204. Based on material property correction parameters, operating condition adaptation parameters, flow uniformity index and distribution stability parameters, coal powder quality index analysis results and pulverizing system optimization parameters are generated through multi-source decision fusion. Among them, the weight allocation of each data source in multi-source decision fusion is dynamically adjusted based on the distribution stability parameters.

[0046] In the coal-fired power industry, the accuracy and real-time performance of pulverized coal quality analysis directly impact system operating efficiency. Related technologies typically rely on single data sources for offline or localized monitoring, resulting in limited evaluation dimensions and difficulty in adapting to dynamic operating conditions. This application proposes a pulverized coal quality data analysis method, aiming to address these issues through a collaborative processing framework of multi-source heterogeneous data. In practical applications, responding to a quality analysis command for pulverized coal within a target pipeline, acquiring multi-source heterogeneous pulverized coal data involves initiating a data acquisition process to collect multi-dimensional information reflecting the physicochemical properties of pulverized coal. This can be achieved by synchronously acquiring real-time signals within the pipeline using a distributed sensor network, or by retrieving pre-stored monitoring records from a historical database. Examples include directly capturing atomic emission spectrometry data using a spectrometer, acquiring three-dimensional holographic reconstruction data using holographic imaging equipment, recording coal flow acoustic data using an acoustic sensor array, measuring spatial charge distribution data using a charge probe array, and extracting dynamic parameters of the pulverizing process from the control system. The main goal is to achieve comprehensive coverage of pulverized coal elemental composition, particle morphology, flow acoustic characteristics, charge behavior, and system operating status. The first type of data processing, which involves processing atomic emission spectroscopy data and 3D holographic reconstruction data to obtain material property correction parameters and particle characterization features, can be understood as eliminating systematic errors from a single data source through data correlation analysis. This can be achieved by using principal component analysis to extract feature vectors from the spectral data and linearly fitting them with the spatial features of the holographic data, or by using a neural network model to perform nonlinear mapping between the two types of data to generate correction parameters. For example, an autoencoder can be used to reconstruct the elemental concentration distribution and combine it with a point cloud clustering algorithm to quantify particle morphology. The main purpose is to achieve a comprehensive characterization of material properties and particle behavior. Furthermore, generating operating condition adaptation parameters based on dynamic parameters of the pulverizing process refers to converting the equipment operating status into quantifiable evaluation indicators. This can be achieved by using a fuzzy logic controller to perform rule-based reasoning on the coal mill current and air pressure parameters, or by using a time series analysis model to extract the trend features of dynamic parameters and map them to operating condition levels. For example, a parameter threshold library can be established based on historical operating data for state classification. This is mainly to achieve dynamic matching between data processing intensity and current operating conditions. Therefore, by performing a second type of data processing on coal flow acoustic data, space charge distribution data, and particle characterization features, we can obtain coal powder flow uniformity index and distribution stability parameters. This can be understood as integrating flow-related data to quantify the coal powder transport state. This can be achieved by using frequency domain analysis to extract the harmonic components of the acoustic signal and cross-validating them with the time-frequency characteristics of the charge signal, or by calculating the covariance matrix of multi-source data based on statistical models to assess flow consistency. For example, we can predict flow state changes and determine stability thresholds using Markov chain models. The main purpose is to achieve objective quantification and stability assessment of flow characteristics.Specifically, based on material property correction parameters, operating condition adaptation parameters, flow uniformity index, and distribution stability parameters, multi-source decision fusion generates coal powder quality index analysis results and pulverizing system optimization parameters. Among these, dynamically adjusting the weight allocation of each data source in multi-source decision fusion based on distribution stability parameters refers to optimizing the input weights of the decision model according to the stability of the flow state. This can be achieved by using the sliding window method to calculate the short-term variance of the distribution stability parameters and adjusting the weight coefficients accordingly, or by setting the mapping relationship between stability level and weight through expert system rules. For example, when the distribution stability parameters are higher than the preset benchmark value, the weight ratio of flow-related data is increased. This is mainly to achieve an adaptive response of the decision-making process to changes in the flow state of coal powder. As a preferred implementation, this embodiment systematically integrates atomic emission spectroscopy data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data, and dynamic parameters of the pulverizing process by constructing a collaborative processing framework for multi-source heterogeneous data, thereby avoiding the limitations of a single data source. At the same time, a dynamic weight allocation mechanism driven by distribution stability parameters is introduced, enabling the multi-source decision fusion process to adjust the contribution of each data source according to the real-time stability of the coal powder flow state, thereby generating optimized parameters that take into account both quality stability and system energy consumption.

[0047] Therefore, this application solves the problems of missing evaluation dimensions, weakened real-time matching ability, and one-sided optimization guidance caused by the single data source and insufficient adaptability to operating conditions in related technologies, and realizes the comprehensiveness of coal powder quality analysis and the improvement of operating condition adaptability.

[0048] In the actual operation of the coal-fired power plant pulverizing system, when a quality analysis command for pulverized coal in the target pipeline is received, the system first collects a multi-source heterogeneous dataset that includes atomic emission spectroscopy data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data, and dynamic parameters of the pulverizing process. This dataset fully covers the elemental composition, particle morphology, flow acoustic characteristics, charge dynamic behavior, and system operating status of pulverized coal, thereby avoiding the limitations of analysis dimensions caused by a single data source.

[0049] In this process, atomic emission spectrometry data and 3D holographic reconstruction data are simultaneously fed into the first type of data processing module. A cross-validation compensation mechanism generates material property correction parameters. Specifically, the particle surface morphology features provided by the 3D holographic reconstruction data are used to inversely correct matrix effect interference in the atomic emission spectrometry data, making the elemental concentration distribution correction more consistent with the actual physical state. Simultaneously, based on the spectral profile parameters of the atomic emission spectrometry data and the spatial distribution density of the 3D holographic reconstruction data, the chemical activity and flow characteristics of the particles are quantitatively characterized, forming particle characterization features. Furthermore, the dynamic parameters of the pulverizing process are analyzed to generate operating condition adaptation parameters. These parameters determine the equipment load state index through the pulverizer current and inlet / outlet pressure difference, and calculate the flow stability index of the pulverizing system by combining the primary air volume and air pressure fluctuation characteristics, enabling the subsequent data processing intensity to dynamically match the current operating conditions. Simultaneously, coal flow acoustic data, space charge distribution data, and the aforementioned particle characterization features are input into the second type of data processing module. The energy distribution characteristics of the coal flow acoustic data in the preset frequency band are corrected for particle morphology parameters to obtain acoustic characteristic parameters. The space charge distribution data is processed using multi-probe signal correlation analysis combined with weighted processing of particle surface characteristic parameters to generate electrostatic characteristic parameters. Finally, based on the consistency index of acoustic and electrostatic characteristic parameters, the coal powder flow uniformity index and distribution stability parameters are determined. Thus, material characteristic correction parameters, operating condition adaptation parameters, flow uniformity index, and distribution stability parameters are sent to the multi-source decision fusion unit. The distribution stability parameter serves as the basis for dynamic weight allocation; when its value is high, the weight of flow-related data is increased, and when it is low, material characteristic data is emphasized, ensuring that the decision-making process adapts to changes in the coal powder flow state. The fusion result, after dynamic adjustment by the operating condition adaptation parameters, outputs the coal powder quality index analysis results, which, together with the distribution stability parameter, are input into a multi-objective optimization model. The model aims to achieve the highest quality index stability and the lowest unit energy consumption of the pulverizing system as collaborative objectives, thereby generating the optimized parameters for the pulverizing system.

[0050] As a specific implementation method, in the pulverizing system of a 600MW coal-fired unit, when a quality analysis command is triggered, the system acquires in real time the atomic emission spectrum data (reflecting the elemental composition distribution), three-dimensional holographic reconstruction data (capturing the three-dimensional morphology of particles), coal flow acoustic data (recording sound wave energy characteristics), space charge distribution data (monitoring charge dynamics), and dynamic parameters of the pulverizing process (including mill current, inlet and outlet pressure difference, primary air volume, etc.) of the pulverized coal in the target pipeline. In the first type of data processing, the atomic emission spectrum data is used to obtain the elemental characteristic intensity distribution after spectral feature analysis, and the three-dimensional holographic reconstruction data is used to extract the particle surface roughness through multi-scale morphology analysis. The two are cross-validated to generate material characteristic correction parameters. Among the dynamic parameters of the pulverizing process, the mill current is 120A and the inlet and outlet pressure difference is 8kPa. Based on this, the equipment load state index is calculated to be 0.75, the primary air volume fluctuation standard deviation is 5%, and the flow stability index of the pulverized coal feeding system is 0.65, forming the operating condition adaptation parameters. In the second type of data processing, the energy distribution characteristics of coal flow acoustic data in the 2-5kHz frequency band were corrected by particle sphericity compensation of 0.85. The space charge distribution data were analyzed by cross-correlation of probe signals combined with surface roughness weighting to obtain a consistency index of 0.82 for acoustic and electrostatic characteristic parameters, which in turn generated a flow uniformity index of 0.78 and a distribution stability parameter of 0.85. In the multi-source decision fusion stage, the weights were dynamically allocated based on the distribution stability parameter of 0.85. The weight of the material property correction parameter was set to 0.4, the weight of the flow uniformity index was set to 0.5, and the weight of the operating condition adaptation parameter was set to 0.1. After fusion, the output coal powder quality index analysis result after adjusting the operating condition adaptation parameter was 0.82. This result was then combined with the distribution stability parameter to optimize the pulverizing system parameters, including adjusting the pulverizer speed to 950 rpm and the primary air volume to 85 m³ / s.

[0051] This technical solution addresses the problem of insufficient evaluation dimensions in related pulverized coal quality analysis methods due to single data sources through systematic collaborative processing of multi-source heterogeneous data. The cross-validation mechanism between material property correction parameters and particle characterization features improves the physical accuracy of element concentration correction and avoids matrix effect interference. Dynamic generation of operating condition adaptation parameters ensures real-time matching of data processing intensity with actual operating conditions, enhancing the adaptability of analysis results. A weight dynamic allocation mechanism driven by distribution stability parameters ensures that the decision fusion process can adapt to changes in pulverized coal flow state, overcoming the deficiency of weak real-time matching capability. The multi-objective optimization model, with quality stability and energy consumption minimization as collaborative objectives, generates optimized parameters for the pulverizing system that simultaneously meet quality requirements and optimal energy efficiency, thus solving the problem of failing to generate optimization guidance that balances quality stability and system energy consumption. Therefore, the comprehensiveness and accuracy of pulverized coal quality assessment are improved, providing reliable technical support for the stable and efficient operation of the pulverizing system.

[0052] In the fields of coal-fired power generation and coal chemical industry, traditional methods for analyzing pulverized coal quality typically rely on single detection methods for offline or localized monitoring, making it difficult to achieve deep fusion of multi-dimensional data and adaptive analysis under operating conditions. Among related technologies, atomic emission spectroscopy data is easily affected by background noise and spectral line overlap, making it impossible to accurately extract elemental characteristics. Three-dimensional holographic reconstruction data lacks refined characterization of the local and global morphology of particles, and no effective correlation is established between the two types of data. This makes material property correction parameters susceptible to distortion due to matrix effects, and particle characterization features cannot comprehensively reflect chemical activity and flow characteristics, ultimately limiting the accuracy of pulverized coal quality analysis and the reliability of system optimization.

[0053] In response, this application further proposes the following technical solution, see [link to technical solution]. Figure 3 Taking the server as the executing entity as an example, the following steps are included.

[0054] 301. Perform spectral feature analysis and background noise suppression on atomic emission spectrum data to obtain the denoised elemental characteristic intensity distribution and spectral profile parameters; 302. Perform multi-scale morphology analysis and spatial distribution reconstruction on the three-dimensional holographic reconstruction data to obtain the surface morphology features and three-dimensional spatial distribution density of the particles; 303. Based on the elemental characteristic intensity distribution and particle surface morphology characteristics, cross-validation compensation processing is performed to obtain material property correction parameters. At the same time, particle characterization features are generated based on spectral profile parameters and three-dimensional spatial distribution density.

[0055] Among them, spectral feature analysis and background noise suppression refer to the techniques for frequency domain decomposition and interference elimination of atomic emission spectral signals. These can be achieved using generalized signal processing techniques such as wavelet transform, Fourier filtering, or adaptive thresholding, with the aim of separating effective elemental feature signals from background noise and avoiding feature loss caused by traditional single denoising methods. Multi-scale morphology analysis and spatial distribution reconstruction refer to the techniques for extracting particle geometric features from different resolution levels and reconstructing spatial distribution. These can be achieved using parallel image processing techniques such as morphological filtering, multi-resolution edge detection, or point cloud clustering, with the aim of simultaneously capturing microscopic details and macroscopic aggregation states of particle surfaces. Cross-validation compensation processing refers to the techniques for mutual verification and error correction using features from different data sources. These can be achieved using generalized mathematical methods such as feature mapping models, regression compensation algorithms, or Bayesian correction frameworks, with the aim of eliminating systematic biases from a single data source. Generating particle characterization features refers to the techniques for fusing multi-dimensional parameters to construct comprehensive characteristic indicators. These can be achieved using parallel data processing techniques such as principal component analysis, feature weighted fusion, or deep feature extraction, with the aim of establishing a correlation between chemical activity and flow characteristics.

[0056] Specifically, the proposed method first performs spectral feature analysis and background noise suppression on atomic emission spectral data. High-frequency noise components are identified and suppressed through frequency domain decomposition, while continuous spectrum background interference is eliminated, thereby obtaining accurate elemental characteristic intensity distribution and spectral profile parameters reflecting the plasma state. Secondly, multi-scale morphology analysis and spatial distribution reconstruction are performed on the three-dimensional holographic reconstruction data. Local features such as microcracks on the particle surface and the overall profile are extracted simultaneously using structural elements at different scales. Then, a three-dimensional spatial distribution density is constructed through point cloud segmentation and feature fusion. Based on this, the theoretical mass absorption coefficient derived from the elemental characteristic intensity distribution and the actual light scattering coefficient determined by the particle surface morphology are cross-validated and compensated to offset measurement deviations caused by matrix effects, generating reliable material property correction parameters. Simultaneously, the half-width at half-maximum (WHM) and peak asymmetry in the spectral profile parameters are converted into a chemical activity index. This is combined with local density extrema and gradient distribution in the three-dimensional spatial distribution density to generate a flow characteristic index, which is then weighted and fused to form a comprehensive particle characterization feature. This technical solution uses a two-way verification mechanism between spectral and morphological data to free elemental feature extraction from background interference, break through the limitations of a single scale in morphological characterization, and establish a physical relationship between chemical properties and spatial distribution, thereby forming a complete data processing closed loop.

[0057] As a specific implementation method, the scheme of this application is implemented as follows: After atomic emission spectral data is acquired by a spectrometer, frequency domain decomposition and background subtraction operations are performed by a digital signal processor; three-dimensional holographic reconstruction data is acquired through a holographic imaging system, and multi-scale edge detection algorithms are used to extract microcracks and sphericity features on the particle surface; subsequently, the element intensity distribution in the spectral data and the surface roughness parameters in the morphology data are input into a feature mapping model to generate a matrix effect correction factor; simultaneously, the peak asymmetry in the spectral profile parameters and the density gradient in the three-dimensional spatial distribution density are combined to calculate the comprehensive characteristic index of the particles through feature weighted fusion. In this embodiment, the spectral data processing module and the morphology analysis module realize data flow through a shared feature database, and the cross-validation compensation unit dynamically adjusts the correction parameters to ensure the physical consistency between the material characteristic correction parameters and the particle characterization features.

[0058] Through the above technical solutions, the accuracy of element feature extraction from atomic emission spectroscopy data is improved, the morphology characterization of three-dimensional holographic reconstruction data covers both microscopic details and macroscopic distribution of particles, the material property correction parameters effectively eliminate matrix effect interference, and the particle characterization features simultaneously reflect chemical activity and flow characteristics, thereby providing an accurate and reliable data foundation for coal powder quality analysis and supporting the execution of subsequent multi-source decision fusion processes.

[0059] Specifically, in some of the embodiments described above in this application, spectral feature analysis and background noise suppression of atomic emission spectral data are proposed to obtain accurate elemental feature information. However, in the implementation process, traditional methods cannot adaptively handle noise interference in different frequency bands, effectively eliminate background interference caused by matrix effects, and separate overlapping spectral lines, resulting in distortion of elemental feature intensity distribution and spectral profile parameter extraction, which in turn affects the accuracy of subsequent material property correction and particle characterization.

[0060] To address this, this application further proposes spectral feature analysis and background noise suppression of atomic emission spectral data to obtain the denoised elemental characteristic intensity distribution and spectral profile parameters, including: Atomic emission spectral data is decomposed into different frequency bands using wavelet packet transform. An adaptive threshold is set based on the energy distribution characteristics of each frequency band, and the high-frequency noise components after decomposition are suppressed based on the corresponding adaptive threshold. The denoised spectral signal is then reconstructed based on the suppressed frequency band components. The denoised spectral signal is subjected to continuous spectrum background subtraction, and an adaptive iterative fitting algorithm is used to eliminate background interference caused by matrix effects, so as to obtain pure elemental characteristic spectral lines. The overlapping spectral lines in the pure element feature spectral lines are deconvolved, and a Gaussian mixture model is used to separate the overlapping spectral lines in the overlapping spectral lines into multiple single-element spectral line components. Multi-peak fitting analysis is performed on multiple single-element spectral line components to obtain the elemental characteristic intensity distribution and spectral profile parameters of each element's characteristic peak. The spectral profile parameters include the full width at half maximum (FWHM), peak asymmetry, and spectral barycenter position.

[0061] Among these, wavelet packet transform processing refers to the technique of multi-resolution decomposition of atomic emission spectral data. It can be implemented using orthogonal or bioorthogonal wavelet bases, aiming to dynamically adjust the noise suppression intensity based on the energy distribution characteristics of each frequency band, avoiding signal detail loss or noise residue caused by a fixed threshold. Adaptive thresholding refers to a noise suppression boundary dynamically set based on the frequency band energy distribution characteristics. It can be determined using energy entropy criteria or variance analysis methods, aiming to make the threshold adapt to signal characteristics, ensuring that high-frequency noise components are suppressed without damaging effective signal components. Continuous spectral background subtraction is a technique for removing continuous background components from the spectrum. It can be implemented using polynomial fitting or spline interpolation methods, aiming to eliminate non-characteristic radiation interference and lay the foundation for subsequent extraction of pure element characteristic spectral lines. Adaptive iterative fitting algorithms refer to dynamically optimizing the fitting process. The computational methods can be implemented using iterative reweighted least squares or Kalman filtering techniques. The aim is to improve the purity of characteristic spectral lines by compensating for nonlinear background interference caused by matrix effects through iterative compensation. Deconvolution processing refers to signal processing techniques for decomposing overlapping spectral line regions. It can be implemented using frequency domain deconvolution or Wiener filtering methods. The aim is to separate mutually interfering spectral line signals into independent components. Gaussian mixture models are spectral line separation methods based on probability distributions. They can be implemented using expectation-maximization algorithms or variational Bayesian inference. The aim is to decompose overlapping spectral lines through the fitting ability of multi-peak probability distributions. Multi-peak fitting analysis is a technique for refined parameter extraction of single-element spectral line components. It can be implemented using nonlinear least squares fitting or Gauss-Newton iterative methods. The aim is to accurately quantify the intensity distribution of elemental characteristics and comprehensively characterize peak features.

[0062] Specifically, the scheme of this application decomposes atomic emission spectrum data into different frequency bands through wavelet packet transform processing. Adaptive thresholds are set using the energy distribution characteristics of each frequency band to specifically suppress high-frequency noise components, thus preserving effective components when reconstructing the denoised spectral signal. Subsequently, continuous spectral background subtraction is performed on the denoised signal to eliminate continuous background interference, and an adaptive iterative fitting algorithm is used to dynamically compensate for nonlinear background interference caused by matrix effects, ensuring the acquisition of pure elemental characteristic spectral lines. Based on this, deconvolution processing is performed on overlapping spectral line regions in the pure elemental characteristic spectral lines. Combining the probability distribution fitting capability of the Gaussian mixture model, the overlapping spectral lines are separated into multiple single-element spectral line components, solving the spectral line overlap problem. Finally, multi-peak fitting analysis is performed on the separated single-element spectral line components to extract elemental characteristic intensity distribution and spectral line contour parameters. This phased and progressive processing flow forms a complete technical chain of noise suppression, background removal, spectral line separation, and parameter extraction. There is a strict execution order and information flow relationship between each step: the output of the previous step serves as the input of the subsequent step, and each step is specifically processed for a particular type of interference. This avoids the shortcomings of traditional methods where a single processing method cannot take into account multiple types of interference, thus systematically ensuring the reliability of element feature information extraction.

[0063] As a preferred embodiment, the specific implementation of the scheme in this application is as follows: In atomic emission spectral data processing, wavelet packet transform is performed using the Symlet wavelet basis to decompose the spectral signal into multiple frequency bands; energy entropy values ​​are calculated based on the energy distribution characteristics of each frequency band, and an adaptive threshold is set accordingly to suppress high-frequency noise components; continuous spectrum background subtraction is achieved using cubic polynomial fitting, and adaptive fitting is performed using iterative reweighted least squares method to eliminate matrix effect interference; frequency domain deconvolution processing is applied to overlapping spectral line regions, and Gaussian mixture model is applied to separate overlapping spectral lines into independent elemental components; finally, nonlinear curve fitting is performed on single-element spectral line components to obtain elemental characteristic intensity distribution and spectral profile parameters such as half-width at half-maximum, peak asymmetry, and spectral centroid position.

[0064] Through the above scheme, this application can adaptively process noise interference in different frequency bands, effectively eliminate background interference caused by matrix effects, and separate overlapping spectral lines, thereby improving the extraction accuracy of elemental characteristic intensity distribution and spectral line profile parameters, and providing highly reliable data support for subsequent material property correction and particle characterization.

[0065] In some of the embodiments described above in this application, multi-scale morphology analysis and spatial distribution reconstruction of three-dimensional holographic reconstruction data are proposed to obtain particle surface morphology features and three-dimensional spatial distribution density. However, in the implementation process, since three-dimensional holographic reconstruction data contains complex particle micro-details and macro-structural information, if only a single-scale processing method is used, it is easy to cause mutual interference or omission in the extraction of local morphology features (such as surface roughness) and global contour features (such as overall geometry), resulting in distortion of surface morphology features and deviation in spatial distribution density calculation, which in turn affects the accuracy of subsequent particle characterization features and ultimately weakens the reliability of coal powder quality analysis.

[0066] In response, this application further proposes to perform multi-scale morphology analysis and spatial distribution reconstruction on the three-dimensional holographic reconstruction data to obtain the particle surface morphology features and three-dimensional spatial distribution density, including: Based on 3D holographic reconstruction data, local morphological features and global contour features of particles are extracted by structural elements of different scales; The surface morphology features of particles are determined based on local morphology features and global contour features; Three-dimensional point cloud segmentation and feature fusion are performed on the surface morphology features of particles to obtain clustered three-dimensional point cloud data. The spatial distribution of the clustered 3D point cloud data is reconstructed to obtain the 3D spatial distribution density, which includes local density extrema and density gradient distribution.

[0067] In practical applications, structural elements at different scales refer to kernel functions used in morphological processing to separate multi-scale features. These can be implemented using circular, square, or custom-shaped structural elements, with sizes ranging from micrometers to millimeters. Their purpose is to adaptively separate the microscopic details and macroscopic structural information of particles through scale differences. Local morphological features refer to the microscopic geometric properties of the particle surface, which can be characterized using parameters such as surface roughness and the depth of minute bumps and depressions. Their purpose is to capture the fine surface state of the particles. Global contour features refer to the overall geometric shape of the particles, which can be described using parameters such as curvature distribution and the radius of curvature of the contour lines. Their purpose is to reflect the macroscopic shape characteristics of the particles. 3D point cloud segmentation refers to the operation of dividing point cloud data into semantically consistent subsets. It can be achieved using density-based clustering algorithms or region growing algorithms, with the aim of eliminating outlier interference and enhancing particle boundary clarity. Feature fusion refers to the process of integrating multi-dimensional geometric attributes. It can be achieved using weighted combinations of attributes such as normal vectors and curvature, with the aim of improving the coherence and signal-to-noise ratio of point cloud data. 3D spatial distribution density refers to the quantitative index of particle aggregation state in space. It can be characterized by local density extrema and density gradient distribution. Local density extrema reflect the location of particle aggregation center, while density gradient distribution represents the rate of density change, with the aim of accurately characterizing the spatial distribution uniformity of particles.

[0068] Specifically, the proposed scheme processes 3D holographic reconstruction data step-by-step using structural elements at different scales. First, it separates local morphological features at the microscale from global contour features at the macroscale, avoiding feature aliasing caused by single-scale processing. Then, it fuses the local and global features in a correlated manner, ensuring that the surface morphology retains high-resolution details while conforming to overall geometric constraints. Based on this, it performs 3D point cloud segmentation and multi-attribute feature fusion on the fused surface morphology features, effectively eliminating noise interference and forming semantically consistent clustered point cloud data. Finally, it reconstructs the spatial distribution based on the optimized point cloud data, accurately quantifying the aggregation state and uniformity of particles in 3D space by calculating local density extrema and density gradient distribution. The entire process, through progressive processing of multi-scale feature extraction, data optimization, and spatial reconstruction, forms a hierarchical feature extraction and verification mechanism, ensuring that particle morphology information from the micro to the macro scale is captured completely and accurately.

[0069] As a preferred embodiment, the solution of this application is implemented as follows: When processing three-dimensional holographic reconstruction data, multiple sets of circular structural elements with diameters of 50 micrometers, 200 micrometers, and 500 micrometers are used to sequentially perform morphological opening operations on the data to extract features at different scales; the point cloud data is segmented using a region growing algorithm, and points with similar curvature are clustered into the same region; feature fusion is performed by integrating normal vectors and curvature information to form semantically consistent point cloud subsets; based on the clustered point cloud, the spatial distribution density is calculated using a kernel density estimation method, local density extrema are identified, and the density gradient change trend is analyzed.

[0070] Through the above scheme, this application solves the problem of imbalance between local and global feature extraction in three-dimensional holographic reconstruction data processing, ensures the acquisition of particle surface morphology features and three-dimensional spatial distribution density, avoids subsequent analysis errors caused by feature distortion, and thus improves the reliability of coal powder quality analysis.

[0071] Specifically, in some embodiments of this application, a cross-validation compensation process based on elemental characteristic intensity distribution and particle surface morphology is proposed to obtain material property correction parameters. However, in its implementation, due to the lack of sufficient consideration of the measurement deviation caused by matrix effect, the theoretical mass absorption coefficient is decoupled from the actual light scattering behavior, resulting in a distorted elemental concentration distribution after correction, which in turn affects the reliability of coal powder quality analysis and the optimization of the pulverizing system.

[0072] In response, this application further proposes to determine the theoretical mass absorption coefficient of each element based on the characteristic intensity distribution of the elements, and to determine the actual light scattering coefficient of the particles based on the surface morphology characteristics of the particles. By comparing and analyzing the theoretical mass absorption coefficient with the actual light scattering coefficient, the matrix effect correction factor is obtained. The elemental characteristic intensity distribution is compensated and corrected based on the matrix effect correction factor to obtain the corrected elemental concentration distribution. Based on the correlation between the corrected elemental concentration distribution and the particle surface morphology, material property correction parameters are generated.

[0073] Among them, the elemental characteristic intensity distribution refers to the intensity distribution of the characteristic peaks of each element in the atomic emission spectrum. It can be obtained by denoising the raw spectral data collected by the spectrometer, and its purpose is to reflect the relative content of each element in the coal powder. The particle surface morphology characteristics refer to the surface physical properties of the coal powder particles, which can include parameters such as surface roughness and sphericity. It can be obtained through three-dimensional holographic reconstruction technology, and its purpose is to characterize the actual influence of particles on light scattering. The theoretical mass absorption coefficient refers to the mass absorption characteristics of an element under standard measurement conditions. It can be determined based on the mapping relationship established by the standard sample calibration experiment, or it can be calculated through the theoretical model of the element's physical properties, and its purpose is to provide a theoretical reference for the intrinsic absorption of elements. The actual light scattering coefficient refers to the scattering effect of particles on light in the actual measurement environment. It can be based on the particle surface... Morphological features are output through a pre-stored mapping table or calculated numerically using optical simulation software, aiming to capture true scattering behavior. Matrix effect correction factors are correction parameters used to quantify the degree of matrix effect interference. They are obtained through deviation analysis of the theoretical mass absorption coefficient and the actual light scattering coefficient, aiming to eliminate measurement distortion caused by matrix effects. Corrected elemental concentration distribution refers to data reflecting the true elemental concentration after compensation correction. It is obtained by inversely correcting the original elemental characteristic intensity distribution through the matrix effect correction factor, aiming to improve the accuracy of concentration analysis. Material property correction parameters are comprehensive correction parameters that integrate chemical composition and physical morphology. They are generated based on the correlation between the corrected elemental concentration distribution and particle surface morphology features, aiming to provide multi-dimensional input for subsequent coal powder quality analysis.

[0074] Specifically, this scheme first determines the theoretical mass absorption coefficient of each element based on its characteristic intensity distribution. Utilizing the inherent correspondence between the characteristic intensity of elements in atomic emission spectra and absorption characteristics under standard measurement conditions, it ensures that the theoretical values ​​accurately reflect the intrinsic properties of the elements. Simultaneously, it determines the actual light scattering coefficient of particles based on their surface morphology. By capturing the actual influence of particles on light scattering through physical parameters such as surface roughness and sphericity, it tightly couples the theoretical model with actual scattering behavior. Comparing and analyzing the theoretical mass absorption coefficient with the actual light scattering coefficient identifies the sources of deviation between the theoretical and actual measurements. This is because matrix effects distort the absorption coefficient by altering the light scattering path; the comparison process directly quantifies the degree of this interference, thereby generating a targeted matrix effect correction factor. Based on this correction factor, the characteristic intensity distribution of elements is compensated and corrected. By inversely correcting systematic errors in the spectral signal, the corrected element concentration distribution accurately reflects the chemical composition of coal powder, avoiding concentration misjudgments caused by neglecting the influence of morphology in traditional methods. Finally, material property correction parameters are generated based on the correlation between the corrected elemental concentration distribution and particle surface morphology. By integrating chemical composition data and physical morphology characteristics, the correction parameters simultaneously reflect the uniformity of elemental distribution and particle structure characteristics, providing a high-fidelity input basis for subsequent coal powder quality analysis.

[0075] As a specific implementation method, the solution of this application is implemented as follows: After acquiring atomic emission spectrum data, the processing unit can query a pre-stored theoretical mass absorption coefficient database based on the elemental characteristic intensity distribution, and simultaneously extract particle surface roughness and sphericity parameters from the three-dimensional holographic reconstruction data; the processing unit can compare the theoretical mass absorption coefficient with the actual light scattering coefficient calculated by optical simulation software to generate a matrix effect correction factor; subsequently, the elemental characteristic intensity distribution is compensated and corrected using this correction factor to obtain the corrected elemental concentration distribution; finally, by combining the corrected concentration distribution with the particle surface morphology characteristics, material property correction parameters are generated for subsequent analysis.

[0076] Through the above-mentioned scheme, this application solves the measurement distortion problem caused by matrix effect, so that the corrected element concentration distribution more accurately reflects the chemical composition of coal powder, avoids the concentration misjudgment caused by ignoring the influence of morphology in traditional methods, and thus improves the reliability of coal powder quality analysis and the optimization of the pulverizing system.

[0077] Specifically, in some embodiments of this application, a method is proposed to determine the theoretical mass absorption coefficient and the actual light scattering coefficient based on the elemental characteristic intensity distribution and particle surface morphology to correct the matrix effect. However, in its implementation, the acquisition of the theoretical mass absorption coefficient does not fully consider the dynamic matching between the element type and the current atomic emission spectroscopy data acquisition conditions, and the determination of the actual light scattering coefficient does not effectively integrate the microscopic details of the particle surface morphology, resulting in a deviation in the calculation of the matrix effect correction factor, which in turn affects the accuracy of the material property correction parameters and ultimately restricts the reliability of coal powder quality analysis.

[0078] In response, this application further proposes determining the theoretical mass absorption coefficient of each element based on the characteristic intensity distribution of elements, and simultaneously determining the actual light scattering coefficient of particles based on the surface morphology characteristics of the particles, including: Based on the element types indicated by the element characteristic intensity distribution, the theoretical mass absorption coefficient of each element under standard measurement conditions consistent with the current atomic emission spectrum data acquisition conditions is obtained by querying the pre-stored first mapping relationship table. The first mapping relationship table is pre-established through standard sample calibration experiments. Surface roughness and sphericity parameters are extracted from the surface morphology features of the particles, and the surface roughness and sphericity parameters are input into a pre-stored second mapping table. The actual light scattering coefficients corresponding to the characteristic wavelengths of atomic emission spectra are output. The second mapping table is pre-established by measuring the light scattering effect at specific wavelengths on standard particle samples with different known morphological features and recording the corresponding relationships.

[0079] The first mapping table is a data structure that stores the correspondence between element types and theoretical mass absorption coefficients. It can be implemented using a relational database or a hash lookup table. Its purpose is to ensure that the theoretical mass absorption coefficient strictly matches the elemental composition and atomic emission spectrum acquisition conditions of the current coal powder. Surface roughness is a parameter that quantifies the degree of microscopic irregularity on the particle surface. It can be implemented using a profilometer or image processing algorithms. Sphericity is a parameter that characterizes the degree to which the particle geometry approximates a sphere. It can be implemented by calculating the shape factor after three-dimensional reconstruction. Its purpose is to describe the microscopic features of the particle surface that affect light scattering behavior. The second mapping table is a data structure that stores the correspondence between surface roughness, sphericity parameters, and actual light scattering coefficients. It can be implemented using a neural network model or a polynomial fitting lookup table. Its purpose is to output the actual light scattering coefficient that matches the particle morphology at a specific atomic emission spectrum characteristic wavelength.

[0080] Specifically, the scheme of this application obtains the theoretical mass absorption coefficient matching the current acquisition conditions by querying the first mapping table based on the element type indicated by the element characteristic intensity distribution. At the same time, it extracts the surface roughness and sphericity parameters from the particle surface morphology features and inputs them into the second mapping table to obtain the actual light scattering coefficient at a specific wavelength. The acquisition of the theoretical mass absorption coefficient and the actual light scattering coefficient provides a reliable basis for matrix effect correction. The first mapping table ensures the dynamic matching of element type and measurement conditions, while the second mapping table realizes the physical correlation between morphological micro-features and light scattering effects. The two work together to make the matrix effect correction factor calculation more accurate, thereby ensuring the reliability of the material property correction parameters.

[0081] As a specific implementation method, the solution of this application is implemented as follows: The first mapping relationship table can be specifically implemented as an SQLite database, which stores the theoretical mass absorption coefficients of different elements under standard measurement conditions; the surface roughness and sphericity parameters are extracted from the surface morphology features of the particles through a three-dimensional point cloud data processing algorithm, the surface roughness is calculated using the root mean square height, and the sphericity is determined by the ratio of the equivalent sphere diameter to the actual volume; the second mapping relationship table is constructed as a polynomial fitting model based on the measured data of the standard particle sample, and the actual light scattering coefficient of the corresponding wavelength is output after inputting the surface roughness and sphericity parameters.

[0082] Through the above technical solutions, the theoretical mass absorption coefficient can be obtained in strict accordance with the elemental composition and atomic emission spectrum acquisition conditions of the current coal powder, avoiding the systematic deviation of the general theoretical value caused by elemental differences or changes in the measurement environment; the determination of the actual light scattering coefficient effectively integrates the microscopic details of the particle surface morphology, overcoming the distortion in scattering coefficient estimation caused by ignoring morphological details in traditional methods; thus improving the calculation accuracy of the matrix effect correction factor, ensuring the accuracy of material property correction parameters, and ultimately improving the reliability of coal powder quality analysis.

[0083] Specifically, in some of the embodiments described above in this application, particle characterization features are generated based on spectral profile parameters and three-dimensional spatial distribution density. However, in the implementation process, the features are generated directly by relying on parameters of only one dimension without fully integrating the dynamic correlation between chemical activity and flow characteristics. Furthermore, the influence of spatial distribution non-uniformity on the characterization of the overall state of the particles is ignored. As a result, the generated particle characterization features cannot accurately reflect the comprehensive behavioral characteristics of coal powder particles under real working conditions, thereby affecting the accuracy of subsequent coal powder quality index analysis and the reliability of pulverizing system optimization.

[0084] In this regard, this application further proposes steps for generating particle characterization features based on spectral profile parameters and three-dimensional spatial distribution density, including: The chemical activity index of the particles is determined based on the full width at half maximum (FWHM) and peak asymmetry in the spectral profile parameters. The flow characteristic index of particles is determined based on the local density extrema and density gradient distribution in the three-dimensional spatial distribution density. The chemical activity index and the flow property index are weighted and fused to obtain the comprehensive particle characterization features; Based on the three-dimensional spatial distribution density, the spatial distribution uniformity of the comprehensive particle characterization features is corrected to obtain the particle characterization features.

[0085] Among them, the half-width at half-maximum (HWHM) refers to the width of the characteristic peak in an atomic emission spectrum when its intensity drops to half of its maximum value. It can be measured using Gaussian fitting or Lorentz fitting algorithms, aiming to quantify the broadening of the energy levels of the excited state of an atom. Peak asymmetry refers to the ratio of the HWHM to the left and right half-maximum of the spectral characteristic peak. It can be calculated using numerical integration or peak fitting methods, aiming to reflect the interference effect of the microstructure of the particle surface on spectral emission. Local density extrema refer to the density peak value of the particle aggregation region in a three-dimensional spatial distribution. It can be determined by kernel density estimation or cluster analysis, aiming to capture the dynamic changes of the particle aggregation region. Density gradient distribution refers to the rate and direction of spatial density change. It can be calculated using the finite difference method or gradient vector flow, aiming to characterize the severity of spatial density transition. Weighted fusion refers to combining the chemical activity index and the flow characteristic index according to a preset weight. It can be implemented using linear weighting or nonlinear neural network models, aiming to balance the contributions of chemical properties and kinematic properties. Spatial distribution uniformity correction refers to correcting the characteristic values ​​according to the actual spatial density distribution. It can be implemented using spatial interpolation or local weighted regression methods, aiming to eliminate feature distortion caused by local aggregation or sparsity.

[0086] Specifically, the proposed method first determines the chemical activity index of particles based on the full width at half maximum (FWHM) and peak asymmetry in the spectral profile parameters. The FWHM directly correlates with the broadening of energy levels in the excited states of atoms, while peak asymmetry reflects the interference effect of the particle surface microstructure on spectral emission. Together, they quantify the potential chemical reaction capability of the particles. Simultaneously, the flow characteristic index of particles is determined based on the local density extrema and density gradient distribution in the three-dimensional spatial distribution density. The local density extrema capture the dynamic changes in the particle aggregation region, while the density gradient distribution characterizes the severity of spatial density transitions, thus realistically reflecting the motion resistance and dispersion trend of pulverized coal in the airflow. Subsequently, the chemical activity index and flow characteristic index are weighted and fused to obtain a comprehensive particle characterization feature, balancing the contributions of chemical and kinematic properties through dynamic weighting. Finally, the comprehensive particle characterization feature is spatially uniformly corrected based on the three-dimensional spatial distribution density. Actual data of the spatial density distribution is used to correct local deviations in the characteristic values, ensuring that the final particle characterization feature represents the uniform state of the overall pulverized coal flow.

[0087] As a specific implementation method, the scheme of this application is implemented as follows: After acquiring atomic emission spectrum data and three-dimensional holographic reconstruction data, the Gauss-Newton algorithm is used to fit the spectral characteristic peaks and calculate the full width at half maximum (FWHM) and peak asymmetry. Simultaneously, an octree data structure is used to spatially partition the three-dimensional point cloud, and the local density extrema and density gradient distribution of each region are calculated. The weighted fusion of the chemical activity index and the flow characteristic index is achieved through an adaptive weight allocation module, which can be specifically an ARM Cortex-M7 microcontroller. Spatial distribution uniformity correction is achieved using an inverse distance weighted interpolation method to locally correct the comprehensive particle characterization features.

[0088] Through the above technical solution, this application can generate particle characterization features that more accurately reflect the comprehensive behavior characteristics of pulverized coal particles under real working conditions, effectively avoiding the problem of separate characterization of chemical activity and flow characteristics, and solving the problem of feature distortion caused by local aggregation or sparseness, thereby improving the accuracy of subsequent pulverized coal quality index analysis and the reliability of pulverizing system optimization.

[0089] Specifically, in some of the embodiments described above in this application, operating condition adaptation parameters are proposed to dynamically adjust the data processing intensity. However, in the process of implementation, the generation of operating condition adaptation parameters lacks a refined quantitative assessment of the key operating states of the pulverizing system. Relying solely on general dynamic parameters makes it difficult to capture the fluctuations in coal mill load and the stability changes of the coal feeding system, resulting in the inability of the data processing intensity to be dynamically adapted to the operating conditions in real time. This, in turn, affects the accuracy of multi-source data fusion and the targeted nature of pulverizing system optimization.

[0090] In response, this application further proposes the following technical solution for generating operating condition adaptation parameters, see [link to relevant documentation]. Figure 4 Taking the server as the executing entity as an example, the following steps are included.

[0091] 401. Extract the operating parameters of the coal mill from the dynamic parameters of the pulverizing process, and determine the equipment load status index based on the coal mill current and inlet / outlet pressure difference in the operating parameters of the coal mill; 402. Extract the powder feeding system parameters from the dynamic parameters of the powder making process, and determine the flow stability index of the powder feeding system based on the primary air volume and air pressure fluctuation characteristics in the powder feeding system parameters; 403. Based on the equipment load state index and the powder feeding system flow stability index, generate operating condition evaluation parameters; 404. Based on the operating condition evaluation parameters, the operating condition adaptation parameters are determined through the operating condition matching model. The operating condition adaptation parameters are used to dynamically adjust the processing intensity of the first type of data processing and the second type of data processing.

[0092] Among them, the equipment load state index is a comprehensive indicator used to quantify the actual working intensity of the coal mill. It can be achieved by weighted combination of coal mill current and inlet / outlet pressure difference or by principal component analysis. The purpose is to avoid load misjudgment caused by single current monitoring through multi-dimensional parameter fusion. The flow stability index of the coal conveying system can be understood as a dynamic quantitative parameter characterizing the continuous stability of the coal powder conveying process. It can be achieved by time-frequency domain joint analysis of primary air volume and air pressure fluctuation characteristics or by statistical feature extraction methods. The purpose is to capture dynamic disturbances of the system rather than static air volume values. The specific operating condition evaluation parameters are the equipment load state index and the flow stability index of the coal conveying system. The fusion results of qualitative indices, for example, can be achieved through multi-source information fusion using the analytic hierarchy process (AHP) or Bayesian networks, with the aim of forming a collaborative assessment of the equipment and system states. The operating condition matching model refers to the conversion mechanism that maps operating condition status assessment parameters into specific adjustment instructions. This can be implemented using a pre-set rule base or a decision tree model trained on historical data, with the aim of transforming assessment results into executable adjustment strategies. Operating condition adaptation parameters can be understood as control parameters used to dynamically adjust the intensity of data processing. These can be implemented using continuously adjustable values ​​or discrete level indicators, with the aim of enabling real-time optimization of data processing intensity as operating conditions change.

[0093] Specifically, the proposed solution operates through a hierarchical quantitative assessment mechanism. First, it determines the equipment load status index based on the pulverizer current and inlet / outlet pressure difference. This index comprehensively reflects the equipment's energy consumption load and material flow resistance, avoiding assessment bias caused by relying solely on current values. Simultaneously, it determines the flow stability index of the pulverizing system based on the primary air volume and air pressure fluctuation characteristics. This index captures the conveying capacity threshold and the degree of dynamic disturbance in the system, and is better at characterizing continuous stability than static air volume monitoring. Subsequently, the two indices are merged to generate operational status assessment parameters, achieving coordinated quantification of key states at the equipment and system levels. Finally, the assessment parameters are transformed into operational status adaptation parameters through an operational condition matching model, dynamically adjusting the processing intensity of the first and second types of data processing to ensure that the multi-source data fusion process always adapts to the current pulverizing state, thereby forming a closed-loop optimization mechanism from operational condition perception to data processing.

[0094] As a preferred embodiment, the specific implementation of this application is as follows: In the pulverizing system of a coal-fired power plant, the current and inlet / outlet pressure difference signals of the coal mill are collected in real time, and the equipment load state index is calculated by an embedded microcontroller; simultaneously, the dynamic signals of the primary air volume sensor and the air pressure sensor are collected, and the air pressure fluctuation characteristics are analyzed using a sliding window to determine the flow stability index of the pulverizing system; the two indices are input into a preset fuzzy logic controller to generate operating condition evaluation parameters; these evaluation parameters are fed into the operating condition matching model, which uses a lookup table method combined with linear interpolation to output operating condition adaptation parameters; the operating condition adaptation parameters are then used to adjust the sampling frequency and filtering intensity of the atomic emission spectroscopy data processing module and the coal flow acoustic data processing module, wherein when the equipment load state index is high, the resolution of the first type of data processing is automatically increased, and when the flow stability index of the pulverizing system is low, the noise suppression capability of the second type of data processing is enhanced.

[0095] Through the above scheme, this application can achieve a refined and quantitative assessment of the key operating status of the pulverizing system, so that the generation of operating condition adaptation parameters can capture the load fluctuations of the coal mill and the stability changes of the pulverizing system, thereby ensuring that the data processing intensity is dynamically adapted to the operating conditions in real time, improving the accuracy of multi-source data fusion and the pertinence of pulverizing system optimization.

[0096] Specifically, in some embodiments of this application, a second type of data processing is proposed to obtain coal powder flow uniformity index and distribution stability parameter by processing coal flow acoustic data, space charge distribution data and particle characterization characteristics. However, in the implementation process, the coal flow acoustic data is affected by particle morphology, resulting in distortion of energy distribution characteristics. The space charge distribution data is affected by particle surface characteristics, resulting in inaccurate cross-correlation coefficient analysis. Furthermore, there is a lack of consistency verification mechanism between acoustic features and electrostatic features. As a result, the generated flow uniformity index and distribution stability parameter cannot truly reflect the coal powder flow state, thereby affecting the reliability of subsequent coal powder quality assessment.

[0097] In response, this application further proposes the following technical solution, see [link to technical solution]. Figure 5 Taking the server as the executing entity as an example, the following steps are included.

[0098] 501. Extract the energy distribution characteristics of coal flow acoustic data in the preset frequency band, and at the same time compensate and correct the energy distribution characteristics based on the particle morphology parameters in the particle characterization characteristics to obtain acoustic feature parameters. 502. Perform multi-probe signal correlation analysis on space charge distribution data to determine the cross-correlation coefficients between different probe signals. Combine the surface characteristic parameters in particle characterization to weight the cross-correlation coefficients and obtain electrostatic characteristic parameters. 503. Based on acoustic characteristic parameters and electrostatic characteristic parameters, determine the consistency index between acoustic characteristic parameters and electrostatic characteristic parameters; 504. When the consistency index exceeds the preset threshold, a coal powder flow uniformity index is generated based on the weighted average of acoustic characteristic parameters and electrostatic characteristic parameters, and the distribution stability parameter is determined based on the variance of the cross-correlation coefficient.

[0099] In practical applications, particle morphology parameters refer to a set of parameters describing the geometric characteristics of pulverized coal particles. These parameters may include particle sphericity, surface roughness, etc., and their purpose is to quantify the influence of particle morphology on the propagation characteristics of sound waves in pulverized coal media, thereby providing a basis for acoustic signal compensation and correction. Surface characteristic parameters can be understood as parameters reflecting the physicochemical properties of the pulverized coal particle surface. These can be implemented using parameters such as surface roughness, conductivity, or dielectric constant. Their purpose is to characterize the degree of interference of particle surface characteristics on charge distribution signals, providing a basis for the weighted processing of cross-correlation coefficients. Specifically, the consistency index is a quantitative indicator used to evaluate the degree of matching between acoustic and electrostatic characteristic parameters. It can be implemented using a weighted fusion value of the time series correlation coefficient after dynamic time warping and the Euclidean distance. Its purpose is to provide a reliable basis for the fusion of acoustic and electrostatic characteristics, avoiding erroneous conclusions caused by the direct fusion of inconsistent features. In practical applications, the distribution stability parameter refers to a comprehensive index that reflects the stability of pulverized coal flow in time and space. It can be achieved by weighted fusion of parameters such as the variance of the cross-correlation coefficient, the extreme value of the time delay, and the dispersion of the time delay. Its purpose is to accurately capture the stability changes of the pulverized coal flow state and provide a reliable basis for subsequent pulverized coal quality assessment.

[0100] Specifically, the proposed method first extracts the energy distribution characteristics of coal flow acoustic data within a preset frequency band. Simultaneously, it uses particle morphology parameters to compensate and correct these energy distribution characteristics, eliminating acoustic signal distortion caused by particle morphology and obtaining accurate acoustic characteristic parameters. Subsequently, multi-probe signal correlation analysis is performed on the space charge distribution data to determine the cross-correlation coefficients between signals from different probes. These cross-correlation coefficients are then weighted using particle surface characteristic parameters to correct signal correlation deviations caused by surface characteristics, resulting in accurate electrostatic characteristic parameters. Based on this, dynamic time warping is used to align the time series of the acoustic and electrostatic characteristic parameters, and the correlation coefficient and Euclidean distance are calculated to determine their consistency index. When the consistency index exceeds a preset threshold, the weights of the acoustic and electrostatic characteristic parameters are dynamically allocated according to the magnitude of the consistency index, and a weighted average is calculated to generate a coal powder flow uniformity index. Simultaneously, the variance of the cross-correlation coefficients is analyzed to determine the flow stability level reflecting time-dimensional fluctuations, and a distribution stability parameter is generated by combining the time delay distribution characteristics. This step-by-step processing and dynamic verification mechanism ensures the accuracy and reliability of the coal powder flow state assessment.

[0101] As a specific implementation method, the scheme of this application is implemented as follows: Acoustic sensors and multiple charge probes are installed on the pulverized coal conveying pipeline. The acoustic sensors collect coal flow acoustic data, and the charge probes collect spatial charge distribution data. The system first performs spectral analysis on the acoustic data to extract energy distribution characteristics of specific frequency bands. Simultaneously, it obtains particle morphology parameters from particle characterization characteristics, calculates the sound wave propagation attenuation coefficient, and performs inverse compensation on the energy distribution characteristics to obtain acoustic characteristic parameters. For the charge probe signals, the system divides the signals of each probe into time windows, determines the cross-correlation coefficients between probe combinations, and weights the cross-correlation coefficients according to particle surface characteristic parameters to obtain electrostatic characteristic parameters. Subsequently, the system performs time series alignment processing on the acoustic characteristic parameters and electrostatic characteristic parameters, calculates correlation indices, and determines consistency indices. When the consistency index exceeds a preset threshold, the system assigns weights according to the consistency index, calculates the weighted average of the acoustic characteristic parameters and electrostatic characteristic parameters, and generates a pulverized coal flow uniformity index. At the same time, it analyzes the statistical characteristics of the cross-correlation coefficients to determine the flow stability level and generates distribution stability parameters in combination with time delay distribution characteristics.

[0102] Through the above scheme, this application can effectively eliminate the interference of particle morphology on acoustic signals, correct the influence of surface characteristics on charge signals, and ensure the accuracy of acoustic and electrostatic characteristic parameters. By introducing a consistency verification mechanism, the direct fusion of inconsistent features is avoided, and the reliability of the coal powder flow uniformity index is improved. At the same time, by using the variance of the cross-correlation coefficient and the time delay distribution characteristics, the stability changes of coal powder flow in the time and space dimensions are captured, so that the distribution stability parameters can truly reflect the coal powder flow state, thereby providing an accurate and reliable decision basis for subsequent coal powder quality assessment.

[0103] Traditional coal powder quality analysis methods typically extract energy distribution features directly from coal flow acoustic data when acquiring acoustic characteristic parameters. However, in the implementation process, the influence of particle morphology parameters (such as sphericity and surface roughness) on sound wave propagation is not fully quantified, resulting in a lack of specificity and effectiveness in compensation and correction. This reduces the reliability of acoustic characteristic parameters and consequently affects the accuracy of coal powder flow uniformity indicators.

[0104] In response, this application further proposes a step-by-step approach to compensate and correct energy distribution characteristics based on particle morphology parameters in particle characterization features, thereby obtaining acoustic characteristic parameters. This step includes: Particle sphericity and surface roughness are extracted from particle morphology parameters, and the propagation attenuation coefficient of sound waves in pulverized coal medium is determined based on particle sphericity and surface roughness. The energy distribution characteristics are inversely compensated based on the propagation attenuation coefficient to obtain the compensated energy distribution; The energy distribution after compensation is compared with the reference energy distribution of the standard pulverized coal sample to obtain the energy distribution similarity. The compensated energy distribution is normalized based on the energy distribution similarity to obtain acoustic feature parameters.

[0105] Among them, particle sphericity refers to the degree to which a particle approximates an ideal sphere. It can be measured using image analysis or laser diffraction methods. Specifically, it can be calculated by using 3D point cloud data to determine the ratio of the particle's circumscribed sphere to its actual volume. Its purpose is to quantify the influence of particle geometry on the sound wave reflection path. Surface roughness can be understood as a statistical characterization of the degree of microscopic unevenness on the particle surface. It can be obtained through scanning electron microscopy image analysis or profilometry measurement. Specifically, root mean square height or arithmetic mean deviation can be used as evaluation indicators. Its purpose is to reflect the scattering intensity of sound waves on the particle surface. Propagation attenuation coefficient is a quantitative parameter of energy attenuation when sound waves propagate in pulverized coal medium. It can be calculated based on a theoretical model established using particle morphology parameters, such as using Mie scattering. The theory is derived by combining particle size distribution, aiming to dynamically characterize the obstruction effect of the actual coal powder microstructure on sound energy transmission. Energy distribution similarity can be understood as the degree of matching between the current coal powder state and the standard reference state. It can be realized by dynamic time warping algorithm or Euclidean distance calculation, and can be quantified by normalized cross-correlation coefficient. Its purpose is to provide an objective evaluation basis for normalization processing. Normalization processing refers to the standardization operation that eliminates dimensional differences and enhances parameter comparability. It can be implemented by minimum-maximum scaling or Z-score standardization method. Specifically, the scaling factor can be dynamically adjusted according to the energy distribution similarity. Its purpose is to ensure that the acoustic feature parameters retain the essence of the original soundprint information and integrate the dual optimization of morphological correction and standard reference.

[0106] The proposed solution first extracts particle sphericity and surface roughness from particle morphology parameters. Based on these morphology parameters, it calculates the propagation attenuation coefficient of sound waves in pulverized coal, enabling the attenuation coefficient to dynamically reflect the microstructural characteristics of actual pulverized coal. Then, it uses this propagation attenuation coefficient to perform inverse compensation on the energy distribution characteristics, specifically eliminating sound energy loss caused by particle irregularities and restoring energy distribution characteristics closer to the true state. Next, it compares the compensated energy distribution with the reference energy distribution of a standard pulverized coal sample, generating an energy distribution similarity score as an objective evaluation benchmark. Finally, based on this similarity score, it performs adaptive normalization processing on the compensated energy distribution, ensuring that acoustic characteristic parameters remain stable when the similarity is high and adaptively scaled when the similarity is low. These steps form a complete compensation and correction chain, with each link closely connected: the extraction of morphology parameters provides the basic input for attenuation coefficient calculation; the dynamic characterization of the attenuation coefficient ensures the targeted nature of inverse compensation; the introduction of standard samples establishes an objective comparison benchmark; and the generation of similarity scores guides the adaptive adjustment of the normalization intensity, thus jointly constructing a quantitative mapping mechanism from microscopic morphology to macroscopic acoustic characteristics.

[0107] As a specific implementation method, the solution of this application is implemented as follows: After deploying acoustic fingerprint sensors in the pulverized coal conveying pipeline to collect coal flow acoustic fingerprint data, particle sphericity and surface roughness are first extracted from particle characterization features. Particle sphericity is calculated through point cloud segmentation of three-dimensional holographic reconstruction data, and surface roughness is determined based on particle surface morphology features analyzed from atomic emission spectroscopy data. Subsequently, particle sphericity and surface roughness are input into a preset propagation attenuation model. This model adopts the Mie scattering theory framework and combines it with a pulverized coal particle size distribution database to output the propagation attenuation coefficient of the sound wave in the current pulverized coal medium. Using this sensor… The attenuation coefficient is used to inversely compensate for the energy distribution characteristics of the acoustic data in the preset frequency band. The compensation process uses a frequency domain deconvolution algorithm to eliminate attenuation distortion caused by shape. The compensated energy distribution is dynamically time-normalized and compared with the reference energy distribution of the standard anthracite powder sample stored in the system to calculate the energy distribution similarity. Finally, the compensated energy distribution is normalized according to the similarity. When the similarity is higher than 0.85, linear scaling is used to maintain the distribution shape. When the similarity is lower than 0.85, a nonlinear scaling factor is introduced to enhance the feature weight of the low frequency band. Finally, the acoustic feature parameters are output for coal powder flow uniformity analysis.

[0108] Through the above technical solutions, this application achieves refined quantification of the influence of particle morphology parameters, enabling the acoustic wave propagation attenuation coefficient to dynamically reflect the microstructural characteristics of actual coal powder, avoiding the distortion of the attenuation model caused by ignoring morphological differences in traditional methods. At the same time, by introducing standard coal powder samples as objective benchmarks, the generation of energy distribution similarity ensures that the correction process has a quantifiable evaluation basis, while the adaptive normalization processing based on similarity effectively eliminates dimensional differences and enhances the comparability of parameters, thereby improving the accuracy and robustness of acoustic characteristic parameters in the calculation of coal powder flow uniformity index.

[0109] Specifically, in some embodiments of this application, a multi-probe signal correlation analysis is proposed to determine the cross-correlation coefficient of space charge distribution data. However, in its implementation, due to the dynamic changes in the coal powder flow state and the susceptibility of probe signals to noise interference, directly calculating the cross-correlation coefficient may result in results that are sensitive to time fluctuations and cannot accurately reflect the spatial distribution characteristics, thereby affecting the reliability of the coal powder flow uniformity index and distribution stability parameters.

[0110] In response, this application further proposes to perform multi-probe signal correlation analysis on the space charge distribution data to determine the cross-correlation coefficients between signals from different probes, including: The charge signals of each probe are divided into multiple time segments by a sliding time window to obtain signal sequences for multiple time periods; Based on the spatial arrangement of multiple probes, determine the probe combinations that need to be correlated and determine the cross-correlation coefficient between the signal sequences of two probes in each probe combination. Determine the peak cross-correlation coefficient and corresponding time delay for each probe combination; Based on the peak cross-correlation coefficients and corresponding time delays of all probe combinations, a probe signal correlation matrix is ​​constructed. The cross-correlation coefficient is determined based on the correlation strength distribution of each probe combination in the probe signal correlation matrix.

[0111] Sliding time window segmentation refers to dividing a continuous charge signal into overlapping or non-overlapping segments according to time sequence. This can be achieved using a fixed-length window or an adaptive-length window, aiming to dynamically capture the instantaneous changes in the signal during pulverized coal flow and avoid distortion caused by ignoring operational fluctuations in the global analysis. Determining the probe combination based on the spatial arrangement of multiple probes refers to selecting the probe pairs to be analyzed based on the physical layout of the probes on the pipe cross-section. This can be achieved using a ring, linear, or grid arrangement of probes, ensuring that the selection of the analysis object is strictly based on the rationality of the probe's physical layout and effectively eliminating interference from probes whose spatial position is irrelevant. The peak value of the cross-correlation coefficient and the corresponding time delay refer to... The point with the highest signal correlation and its time offset can be calculated using cross-correlation function or phase difference analysis. The purpose is to quantify the maximum correlation between signals and reflect the velocity differences and directional characteristics of pulverized coal flow. The probe signal correlation matrix is ​​a data structure that structurally represents the correlation characteristics of all probe combinations. It can be implemented using a two-dimensional matrix or graph structure. The purpose is to achieve structured integration of multi-dimensional signal relationships, facilitating global stability assessment. Determining the cross-correlation coefficient based on the correlation intensity distribution refers to dynamically adjusting the weights based on statistical laws. It can be implemented using Gaussian distribution fitting or cluster analysis. The purpose is to suppress the influence of noise interference and abnormal signals by analyzing the statistical laws of the intensity distribution.

[0112] Specifically, the proposed solution dynamically captures the instantaneous changes in the signal during pulverized coal flow by dividing the charge signal of each probe into sliding time windows, thus avoiding distortion caused by neglecting operating condition fluctuations in global analysis. Based on the spatial arrangement of multiple probes, the probe combinations for correlation analysis are determined, ensuring that the selection of analysis objects strictly conforms to the rationality of the probe's physical layout and effectively eliminating interference from probes whose spatial location is irrelevant. By determining the peak value of the cross-correlation coefficient and the corresponding time delay for each probe combination, not only is the maximum correlation between signals quantified, but the time delay information also reflects the velocity differences and directional characteristics of pulverized coal flow. A probe signal correlation matrix is ​​constructed based on the peak value of the cross-correlation coefficient and the corresponding time delay of all probe combinations, achieving structured integration of multi-dimensional signal relationships and facilitating global stability assessment. Finally, the cross-correlation coefficient is determined according to the correlation strength distribution of each probe combination in the probe signal correlation matrix. By analyzing the statistical laws of the intensity distribution, the weights are dynamically adjusted to effectively suppress the influence of noise interference and abnormal signals, thereby laying a data foundation for the generation of pulverized coal flow uniformity indicators and distribution stability parameters.

[0113] As a preferred embodiment, the specific implementation of this application is as follows: Eight charge probes are uniformly arranged on the cross-section of the pulverized coal pipeline to form a ring array; the collected charge signals are divided using a sliding time window, and the window length and step size are dynamically adjusted according to the pulverized coal flow velocity; based on the ring arrangement of the probes, adjacent probe pairs are determined as combinations for correlation analysis; the cross-correlation function is calculated for the signal sequence of each probe combination to obtain the peak value of the correlation coefficient and its corresponding time delay; the peak value and time delay data of all probe combinations are organized into a correlation matrix; by analyzing the distribution characteristics of the correlation intensity in the matrix, the final cross-correlation coefficient is determined using a weighted average method, wherein the weights are dynamically adjusted according to the dispersion of the intensity distribution.

[0114] Through the above solution, this application solves the problem of inaccurate correlation coefficient calculation caused by dynamic changes in the flow state of pulverized coal and noise interference from probe signals, so that the cross-correlation coefficient can more reliably reflect the spatial distribution characteristics of pulverized coal flow, thereby improving the accuracy of pulverized coal flow uniformity index and distribution stability parameters.

[0115] In some of the embodiments described above in this application, a consistency index is proposed to be determined based on acoustic characteristic parameters and electrostatic characteristic parameters to evaluate the flow state of pulverized coal. However, in its implementation, since the acoustic signal and space charge signal come from sensors with different physical principles and there is a time series asynchronous problem, directly calculating the correlation will lead to the distortion of the consistency index, thereby affecting the reliability of the pulverized coal flow uniformity index.

[0116] In response, this application further proposes a consistency index for determining the acoustic and electrostatic characteristic parameters based on acoustic and electrostatic characteristic parameters, including: The acoustic and electrostatic characteristic parameters are dynamically time-warped to align them in time series, resulting in aligned acoustic and electrostatic characteristic parameter sequences. Determine the correlation coefficient and Euclidean distance between the aligned acoustic feature parameter sequence and the electrostatic feature parameter sequence; Consistency indices are determined by weighted fusion based on correlation coefficients and Euclidean distance.

[0117] Among them, dynamic time warping refers to an algorithmic framework for flexibly matching non-uniform time series. It can be implemented using dynamic time warping algorithms or their improved variants. Its purpose is to solve the time series misalignment problem caused by differences in sampling rates or signal transmission delays in sensors based on different physical principles. The correlation coefficient can be understood as a statistical measure characterizing the strength of the linear correlation between the changing trends of two time series. It can be implemented using Pearson correlation coefficient or Spearman rank correlation coefficient. Its purpose is to quantify the consistency of acoustic and electrostatic signals in terms of wave patterns. Euclidean distance specifically refers to the geometric distance measure between two time series in multidimensional space. It can be implemented by calculating the square root of the sum of squares of the amplitude differences at corresponding time points. Its purpose is to capture the absolute amplitude deviation between signals. Weighted fusion can be understood as a method of combining multidimensional indicators into a comprehensive evaluation index. It can be implemented using linear weighting functions or nonlinear fusion models. Its purpose is to balance the contributions of trend correlation and amplitude differences to generate a robust consistency index.

[0118] Specifically, the proposed solution first uses dynamic time warping to elastically align the acoustic and electrostatic characteristic parameters over time, eliminating signal time offsets caused by differences in sensor physical characteristics and ensuring that subsequent comparisons are based on a synchronized time reference. Based on this, the correlation coefficient of the aligned sequences is calculated to assess the similarity of signal change trends, while the Euclidean distance is calculated to quantify the absolute difference in signal amplitude; these two dimensions form complementary measurement dimensions. Finally, a weighted fusion mechanism integrates the trend consistency reflected by the correlation coefficient with the amplitude matching degree represented by the Euclidean distance to generate a comprehensive consistency index that fully characterizes the matching degree between acoustic and electrostatic signals. This processing flow eliminates asynchronous interference through time alignment and combines complementary measurements with dual-dimensional indicators to ultimately achieve a more accurate assessment of the coal powder flow state.

[0119] As a specific implementation method, the solution of this application is implemented as follows: In the pulverized coal pipeline monitoring system, the acoustic feature parameter sequence output by the acoustic sensor and the electrostatic feature parameter sequence acquired by the multi-probe charge detection device are input into a dynamic time warping module. This module uses a minimum cumulative distance path search algorithm to achieve elastic signal matching, and the window size is dynamically adjusted according to the pulverized coal flow rate. The aligned sequence is then analyzed by a statistical analysis unit to calculate the Pearson correlation coefficient and the Euclidean distance at the corresponding time point. The correlation coefficient weight is set to 0.65 according to the current operating conditions, and the Euclidean distance weight is set to 0.35. A consistency index is generated through linear weighted fusion. In this implementation method, the dynamic time warping process uses a quadratic penalty function to optimize path continuity, ensuring that signal alignment accuracy is maintained even when the pulverized coal flow rate fluctuates.

[0120] Through the above technical solution, this application solves the problem of consistency index distortion caused by asynchronous time series of acoustic signal and space charge signal, improves the accuracy of coal powder flow state assessment, and thus provides a reliable basis for calculating coal powder flow uniformity index for pulverizing system.

[0121] In some of the embodiments described above in this application, a method for generating pulverized coal flow uniformity index and distribution stability parameter based on acoustic and electrostatic characteristic parameters is proposed. However, in the implementation process, the weight allocation is not dynamically adjusted according to the consistency index, which leads to a decrease in the reliability of the index when the consistency of the acoustic and electrostatic characteristic data sources is low. Furthermore, the distribution stability parameter is determined only based on the variance value of the cross-correlation coefficient, failing to fully integrate the time delay distribution characteristics to reflect the stability of pulverized coal flow in the spatial dimension.

[0122] To address this, this application further proposes generating the coal powder flow uniformity index based on the weighted average of the acoustic characteristic parameters and the electrostatic characteristic parameters, and determining the distribution stability parameter based on the variance of the cross-correlation coefficient, including: Based on the magnitude of the consistency index, the weights of the acoustic feature parameters and the electrostatic feature parameters in the weighted average calculation are determined, wherein feature parameters with higher consistency indices are assigned larger weights. Based on the determined weights, the weighted average of the acoustic characteristic parameters and the electrostatic characteristic parameters is determined, and the weighted average is mapped to a preset flow uniformity level to obtain the coal powder flow uniformity index. Determine the variance of the peak cross-correlation coefficient of all probe combinations within a preset time period, and determine the flow stability level based on the magnitude of the variance. The flow stability level is used to reflect the stability in the time dimension. The distribution stability parameter is generated based on the time delay distribution characteristics of the flow stability level and the peak value of the cross-correlation coefficient.

[0123] Among them, the consistency index refers to the technical parameter used to quantify the matching degree between acoustic and electrostatic characteristic parameters. It can be achieved by weighting the correlation coefficient and Euclidean distance after dynamic time warping, with the aim of evaluating the reliability of multi-source data sources. The weight allocation mechanism of acoustic and electrostatic characteristic parameters refers to the technical means of dynamically adjusting the contribution based on the consistency of data sources. It can be achieved based on threshold comparison or continuous function mapping, with the aim of avoiding index distortion caused by the anomaly of a single data source. The flow uniformity level refers to the technical framework for normalizing the weighted average into a standardized classification system. It can be achieved by discrete level division or continuous interval mapping, with the aim of establishing a direct correlation between the index and the actual operating conditions. The peak variance of the cross-correlation coefficient refers to the technical indicator that quantifies the dispersion of signal fluctuations in the time dimension. It can be achieved through statistical variance calculation or sliding window analysis, with the aim of capturing the stability changes of pulverized coal flow on the time axis. The time delay distribution characteristic refers to the set of technical parameters characterizing the signal time delay characteristics between different probes. It can be achieved based on time delay extreme value extraction or probability distribution modeling, with the aim of reflecting the uniformity of pulverized coal distribution in the pipeline space.

[0124] Specifically, the proposed solution employs a consistency index-driven weight allocation mechanism. When the consistency index between acoustic and electrostatic characteristic parameters is high, a more reliable data source is assigned a larger weight, ensuring the robustness of the flow uniformity index. Based on a weighted average value mapped to a preset flow uniformity level, the standardized quantification of the pulverized coal flow state is achieved. Simultaneously, the flow stability level in the time dimension is determined by the peak variance of the cross-correlation coefficient, and a distribution stability parameter is generated by integrating time delay distribution characteristics, enabling this parameter to simultaneously characterize the stability state in both the time and spatial dimensions. These technical features form an organic whole: the consistency index serves as the decision-making basis for weight allocation, the flow uniformity index reflects the uniformity of pulverized coal flow, and the distribution stability parameter comprehensively evaluates the spatiotemporal stability of pulverized coal flow through the synergistic analysis of the time dimension variance and the spatial dimension time delay characteristics.

[0125] As a specific implementation method, the scheme of this application is implemented as follows: The consistency index can be calculated by using a dynamic time warping algorithm to align the acoustic feature parameter sequence and the electrostatic feature parameter sequence in time. Then, the Pearson correlation coefficient between the sequences is calculated and combined with the normalized Euclidean distance. The consistency index value is obtained by fusing through preset weights. The weight allocation process can set a threshold range for the consistency index. When the index value is in the high consistency range, the weight of the acoustic feature parameter can be increased to above 0.7. The flow uniformity level mapping can divide the weighted average into five discrete levels, corresponding to uniform flow, basically uniform, critically uniform, locally non-uniform and severely non-uniform states, respectively. The extraction of time delay distribution characteristics can analyze the time delay probability distribution of the cross-correlation coefficient peaks, determine the distribution morphology characteristics through kernel density estimation, and thus characterize the spatial distribution characteristics of pulverized coal on the pipeline cross-section.

[0126] Through the above technical solutions, this application solves the problem of decreased reliability of indicators when the consistency of acoustic and electrostatic characteristic data sources is low. The robustness of the flow uniformity index is ensured by the dynamic weight allocation mechanism. At the same time, the distribution stability parameter achieves a comprehensive characterization of the stability of pulverized coal flow in both time and space dimensions by integrating the variance value in the time dimension and the time delay distribution characteristics in the spatial dimension. This avoids the technical defect that cannot be evaluated by relying solely on the variance value, and enables the pulverized coal flow state evaluation results to more accurately reflect the actual working conditions.

[0127] In some of the embodiments described above in this application, a distribution stability parameter is generated based on the time delay distribution characteristics of the flow stability level and the peak value of the cross-correlation coefficient. However, in its implementation, the stability in the spatial dimension is not effectively quantified, and the flow stability level in the time dimension is relied upon only. As a result, the distribution stability parameter cannot fully reflect the spatial distribution characteristics of the coal powder flow, which makes the assessment of the coal powder flow state one-sided, thereby affecting the accuracy of subsequent multi-source decision fusion and the reliability of the pulverizing system optimization.

[0128] In response, this application further proposes to generate distribution stability parameters based on the time delay distribution characteristics of the flow stability level and the peak value of the cross-correlation coefficient, including: The time delay extremes and time delay dispersion in the time delay distribution characteristics of the cross-correlation coefficient peak are determined. The time delay extremes are used to reflect the fluctuation range of the coal powder flow velocity, and the time delay dispersion is used to reflect the degree of difference in flow velocity between different probe combinations. Based on the time delay extreme value and time delay dispersion, the time delay stability coefficient is determined. The increase of the time delay extreme value and / or the increase of the time delay dispersion will lead to a decrease in the time delay stability coefficient. The time delay stability coefficient reflects the stability in the spatial dimension. The distribution stability parameters are obtained by weighting and fusing the flow stability level and the time delay stability coefficient.

[0129] Among them, the time delay extreme value refers to the difference between the maximum and minimum values ​​in the peak time delay sequence of the cross-correlation coefficient. It can be realized by range calculation or dynamic range statistical methods, and its purpose is to quantify the fluctuation range of coal powder flow velocity on the time axis. The time delay dispersion can be understood as the dispersion index of time delay values ​​among different probe combinations. It can be characterized by statistical quantities such as standard deviation, variance, or interquartile range, and its purpose is to capture the non-uniformity of coal powder flow in spatial distribution. The time delay stability coefficient refers to the parameter that maps the time delay extreme value and time delay dispersion to continuous values ​​through a preset functional relationship. It can be realized by linear decay function or exponential decay function, and its purpose is to transform spatial non-uniformity into a quantifiable stability index. Weighted fusion is specifically the operation of integrating stability indices in the time dimension and spatial dimension. It can be realized by fixed weight coefficient or adaptive weight allocation strategy, and its purpose is to construct a distribution stability parameter that comprehensively reflects spatiotemporal characteristics.

[0130] Specifically, the proposed solution analyzes the time delay distribution characteristics of the cross-correlation coefficient peaks. First, it extracts the time delay extremes to characterize the temporal fluctuation range of pulverized coal flow velocity, and simultaneously extracts the time delay dispersion to quantify the spatial difference in flow velocity between different probe combinations. Together, these two methods transform the abstract time delay distribution into a quantifiable spatial non-uniformity index. Subsequently, a time delay stability coefficient is constructed based on the time delay extremes and time delay dispersion. By setting a mapping rule that increases in time delay extremes or time delay dispersion lead to a decrease in the coefficient, spatial non-uniformity is transformed into a continuously adjustable stability index, which clearly characterizes the stability level in the spatial dimension. Finally, the flow stability level, which reflects stability in the time dimension, and the time delay stability coefficient, which reflects stability in the spatial dimension, are weighted and fused. This allows the distribution stability parameter to integrate information in both the spatiotemporal dimensions, avoiding the limitations of single-dimensional evaluation and thus providing a more comprehensive parameter basis for multi-source decision fusion.

[0131] As a specific implementation method, the scheme of this application is implemented as follows: In the pulverized coal pipeline monitoring system, spatial charge distribution data is collected by a multi-probe charge sensor array. After dividing the signals of each probe into sliding time windows, the peak value of the cross-correlation coefficient and the time delay value are calculated. The extreme value of the time delay is determined by the range calculation method, which reflects the maximum fluctuation range of the pulverized coal flow velocity. The time delay dispersion is determined by the standard deviation calculation method, which characterizes the spatial difference of the flow velocity between different probe combinations. Based on the extreme value of the time delay and the time delay dispersion, the time delay stability coefficient is calculated by an exponential decay function, wherein the function parameters are calibrated in advance by flow experiments of standard pulverized coal samples. The flow stability level determined based on the variance of the cross-correlation coefficient and the calculated time delay stability coefficient are linearly weighted and fused according to a preset ratio, and the distributed stability parameter is output for subsequent decision-making process.

[0132] Through the above technical solution, this application achieves a comprehensive evaluation of the spatiotemporal characteristics of pulverized coal flow, solves the problem that the distribution stability parameters cannot fully reflect the spatial distribution characteristics of pulverized coal flow, makes the evaluation results of pulverized coal flow state more accurate and reliable, and thus improves the accuracy of multi-source decision fusion and the reliability of pulverizing system optimization.

[0133] In the fields of coal-fired power generation and coal chemical industry, traditional coal powder quality analysis methods typically rely on single detection methods for offline analysis or local monitoring, such as judging the flow state solely through spectral analysis of elemental composition or charge signals. Such methods struggle to provide a comprehensive coal powder quality assessment that matches real-time operating conditions, leading to the pulverizing system often operating in a suboptimal state, hindering energy efficiency improvement and stable production control. In some embodiments of this application, a method based on dynamically adjusting the weight allocation of each data source in multi-source decision fusion based on distributed stability parameters is proposed to improve the reliability of coal powder quality analysis. However, in its implementation, the weight allocation scheme lacks refined consideration of the magnitude and dynamic trend of distributed stability parameters, resulting in insufficient matching between the weight allocation and the actual reliability of the data sources; the weighted fusion process fails to effectively combine operating condition adaptation parameters for real-time dynamic adjustment, making the fusion results unable to adapt to fluctuations in the operating conditions of the pulverizing system; and the generation of pulverizing system optimization parameters lacks a collaborative optimization mechanism between quality index stability and system energy consumption, making it difficult to simultaneously achieve a balance between coal powder quality stability and optimal pulverizing energy consumption.

[0134] In response, this application further proposes the following technical solution, see [link to technical solution]. Figure 6 Taking the server as the executing entity as an example, the following steps are included.

[0135] 601. Based on the distribution stability parameter, the weights of material property correction parameters, working condition adaptation parameters and flow uniformity index in decision fusion are dynamically allocated to obtain a weight allocation scheme, wherein the distribution stability parameter is positively correlated with the weight of the corresponding data source. 602. Based on the weighting scheme, the material property correction parameters and flow uniformity index are weighted and fused, and the working condition adaptation parameters are used to dynamically adjust the weighting and fusion process to obtain the coal powder quality index analysis results. 603. Input the coal powder quality index analysis results and distribution stability parameters into the multi-objective optimization model, and take the highest quality index stability and the lowest unit energy consumption of the pulverizing system as the synergistic optimization objectives to obtain the optimization parameters of the pulverizing system.

[0136] In practical applications, the distributed stability parameter refers to a quantitative index reflecting the comprehensive stability of pulverized coal flow in both time and space dimensions. It can be implemented using methods such as acoustic signal spectral stability analysis or charge distribution time-domain feature extraction. Its purpose is to objectively assess the reliability of pulverized coal flow state and provide a dynamic weight allocation basis for multi-source data fusion. The weight allocation scheme can be understood as a mathematical transformation mechanism that maps the distributed stability parameter to the contribution of data sources. It can be implemented using nonlinear functions or fuzzy logic rules, aiming to ensure that highly stable data sources receive higher weights during the fusion process, thereby improving the accuracy of the analysis results. The dynamic adjustment of the operating condition adaptation parameter in the weighted fusion process refers to a mechanism that compensates the fusion results based on the real-time operating status of the pulverizing system. It can be implemented based on a combination of equipment load state index and flow stability index, aiming to eliminate the interference of operating condition fluctuations on pulverized coal quality assessment. The multi-objective optimization model can be understood as a mathematical solution framework that simultaneously optimizes the stability of quality indicators and system energy consumption. It can be implemented using genetic algorithms or simulated annealing algorithms, aiming to generate operating parameters that balance pulverized coal quality stability and energy efficiency.

[0137] Specifically, the proposed scheme first dynamically determines the weighting scheme of material property correction parameters, operating condition adaptation parameters, and flow uniformity index based on the numerical characteristics of the distribution stability parameters. Since the distribution stability parameters directly characterize the spatiotemporal stability of pulverized coal flow, a higher value indicates a stronger reliability of the corresponding data source. Therefore, a positive correlation is used to ensure that high-stability data sources receive higher weights. Subsequently, the material property correction parameters and flow uniformity index are weighted and fused according to the weighting scheme. At the same time, the operating condition adaptation parameters are introduced to dynamically adjust the fusion process. The equipment load state index is used to adjust the benchmark value of the fusion result, and the pulverized coal feeding system flow stability index is used to suppress result fluctuations. Finally, the pulverized coal quality index analysis results obtained by fusion are input together with the distribution stability parameters into a multi-objective optimization model. Under the constraint that the distribution stability parameters are not lower than a preset threshold, the optimal pulverizing system operating parameter vector is generated by synergistically optimizing the stability of the quality index and the unit energy consumption.

[0138] As a preferred embodiment, the solution of this application is implemented as follows: When the distribution stability parameter shows that the pulverized coal flow is in a highly stable state through cross-correlation analysis, the system automatically increases the proportion of the material characteristic correction parameter in the weight allocation scheme; when the operating condition adaptation parameter detects that the pulverizer is running at high load, the dynamic adjustment mechanism applies a positive baseline adjustment to the weighted fusion result; after receiving the pulverized coal quality index analysis result, the multi-objective optimization model uses the particle swarm optimization algorithm to search for a balance point between quality stability and energy consumption objectives, and the output pulverizing system optimization parameters include pulverizer speed adjustment commands and air volume ratio adjustment schemes. In this embodiment, the processing unit can be specifically an industrial-grade embedded controller, the data fusion module realizes dynamic weight calculation through software algorithms, and the optimization model solution process is completed in distributed computing nodes.

[0139] Through the above technical solution, this application achieves the matching of weight allocation and data source reliability, enabling the coal powder quality analysis results to adapt to the fluctuations in the pulverizing system's operating conditions. At the same time, while ensuring the stability of coal powder quality, it reduces the unit energy consumption of the pulverizing system, solving the technical problems of inaccurate weights, poor adaptability to operating conditions, and single optimization objective in multi-source decision fusion.

[0140] In practical applications, some embodiments of this application propose to dynamically adjust the weight allocation of each data source in multi-source decision fusion based on the distribution stability parameter. However, in its implementation, the weight allocation is based only on the static value of the distribution stability parameter, and the dynamic characteristics of the parameter changing over time are not fully considered. This results in the weight adjustment failing to respond to fluctuations in operating conditions in a timely manner, and may introduce decision bias due to sudden changes in the stability of the data source, affecting the accuracy of the coal powder quality analysis results and the reliability of the system optimization.

[0141] In this regard, this application further proposes the following steps for a weight allocation scheme: Based on the distribution stability parameter, the weights of material property correction parameters, operating condition adaptation parameters, and flow uniformity index are dynamically allocated in the decision fusion, resulting in a weight allocation scheme, including: Based on the magnitude of the distribution stability parameter, initial weights are assigned to the material property correction parameter, the working condition adaptation parameter, and the flow uniformity index, respectively. Among them, data sources with higher distribution stability parameters are assigned higher initial weights. The distribution stability parameter is obtained as a trend within a preset time window. When the trend shows that the parameter value is stable, the weight of the corresponding data source is increased. When the trend shows that the parameter value fluctuates drastically, the weight of the corresponding data source is decreased. The weights adjusted for trend characteristics are normalized so that the sum of the weights of the material property correction parameters, operating condition adaptation parameters, and flow uniformity index is 1, thus obtaining the weight allocation scheme.

[0142] Specifically, initial weight setting refers to quantifying the reliability level of the data source based on the instantaneous value of the distribution stability parameter. This can be achieved using linear proportional mapping or piecewise threshold judgment, aiming to ensure that highly stable data sources dominate in the initial stage of fusion and avoid low-quality data interfering with the decision-making process. Trend feature acquisition refers to analyzing the dynamic evolution of the distribution stability parameter over time, which can be achieved through sliding window statistical analysis or trend line fitting algorithms. The aim is to identify the stable or fluctuating state of the parameter, providing a basis for dynamic weight adjustment. In practical applications, dynamic weight adjustment refers to real-time correction of the initial weights based on the trend features. This can be achieved using gain factor adjustment or rule engine triggering mechanisms, aiming to strengthen the contribution of reliable data sources when parameters are continuously stable and suppress the impact of sudden operational disturbances when parameters fluctuate drastically. Specifically, normalization processing refers to standardizing the adjusted weights, which can be achieved using linear scaling or probabilistic normalization methods. The aim is to ensure that the sum of the weights is strictly 1, ensuring the mathematical rigor and interpretability of the fusion results.

[0143] Specifically, the proposed scheme first sets initial weights based on the current values ​​of the distributed stability parameters, giving higher weight proportions to high-stability data sources. Then, by analyzing the changing trends of the parameters within a preset time window, the weights of each data source are dynamically adjusted. When parameter values ​​remain stable, the corresponding weights are increased to enhance the contribution of reliable data; when parameter values ​​fluctuate drastically, the weights are decreased to suppress interference. Finally, the adjusted weights are normalized so that the sum of the three weights is always 1, thereby generating a standardized weight allocation scheme. This mechanism, by integrating the instantaneous values ​​and temporal evolution characteristics of the distributed stability parameters, constructs a refined dynamic weight allocation logic, enabling weight adjustments to respond in real time to stability changes during pulverized coal transportation, effectively maintaining the adaptability and robustness of multi-source decision fusion under complex operating conditions.

[0144] As a specific implementation method, the scheme of this application is implemented as follows: During the monitoring of pulverized coal transportation, the system collects distribution stability parameters in real time. When the parameter is detected to fluctuate smoothly within a continuous time window, the weight ratio of the material characteristic correction parameter is automatically increased. If the parameter fluctuates, the weight of the working condition adaptation parameter is reduced accordingly. After the weight adjustment, the normalization algorithm is used to ensure that the sum of the three weights is always 1, thereby generating a dynamically adaptive weight allocation scheme to provide standardized input for subsequent pulverized coal quality index analysis.

[0145] Through the above technical solutions, weight allocation can respond promptly to fluctuations in operating conditions, reduce decision-making biases caused by sudden changes in data source stability, and improve the accuracy of pulverized coal quality analysis results and the reliability of pulverizing system optimization.

[0146] Specifically, in some of the embodiments described above in this application, it is proposed to dynamically adjust the weighted fusion process using operating condition adaptation parameters to generate coal powder quality index analysis results. However, in the implementation process, since the fluctuations in equipment load status and the changes in the flow stability of the coal feeding system are not incorporated into the adjustment mechanism of the fusion results in real time, the weighted fusion results show baseline drift and increased fluctuations. This makes it impossible for the coal powder quality assessment to accurately reflect the true quality status when the operating conditions change dynamically, thereby affecting the reliability of subsequent pulverizing system optimization.

[0147] In response, this application further proposes to dynamically adjust the weighted fusion process using operating condition adaptation parameters to obtain coal powder quality index analysis results, including: Based on the equipment load state index in the operating condition adaptation parameters, the baseline adjustment amount of the weighted fusion result is determined. The higher the equipment load state index, the larger the baseline adjustment amount. Based on the flow stability index of the powder feeding system in the working condition adaptation parameters, the fluctuation compensation factor of the weighted fusion result is determined. The lower the flow stability index, the larger the fluctuation compensation factor to suppress the fluctuation of the fusion result. The material property correction parameters and flow uniformity index are weighted and summed according to their weights to obtain the initial fusion value. The initial fusion value is then dynamically corrected using the baseline adjustment amount and fluctuation compensation factor to obtain the coal powder quality index analysis results.

[0148] Among them, the equipment load state index is a quantitative indicator reflecting the operating intensity of the coal mill. It can be realized by using the normalized combination of the coal mill current and the inlet and outlet pressure difference. Its purpose is to transform the equipment operating status into a quantifiable adjustment basis. The baseline adjustment can be understood as the benchmark correction value of the weighted fusion result. It can be determined by using a relationship function that is linearly or nonlinearly positively correlated with the equipment load state index. Its purpose is to compensate for the systematic shift in pulverized coal characteristics caused by the increased operating intensity of the coal mill under high load conditions. The pulverized coal feeding system flow stability index is a quantitative indicator reflecting the stability of the pulverized coal feeding system. It can be realized by using the weighted combination of the primary air volume fluctuation amplitude and the air pressure standard deviation. Its purpose is to characterize the smooth operation of the pulverized coal feeding system. Stability; the fluctuation compensation factor can be understood as a dynamic adjustment factor that suppresses fluctuations in the fusion result. It can be determined using an exponential function or piecewise function that is negatively correlated with the flow stability index of the coal feeding system. Its purpose is to smooth out instantaneous noise interference caused by air volume fluctuations or pipeline blockages. The initial fusion value refers to the intermediate calculation result obtained by weighted summation of material characteristic correction parameters and flow uniformity index. Its purpose is to integrate the core information of coal powder material composition and flow characteristics. Dynamic correction refers to the application of baseline adjustment and fluctuation compensation factor to the calculation process of the initial fusion value. It can be achieved by combining additive compensation and multiplicative attenuation. Its purpose is to make the final analysis result reflect both the essence of coal powder quality and adapt to changes in the operating environment.

[0149] Specifically, the proposed solution maps the equipment load state index and the pulverized coal feeding system flow stability index in the operating condition adaptation parameters to a baseline adjustment amount and a fluctuation compensation factor, respectively, achieving dual dynamic adjustment of the weighted fusion process. When the equipment load state index increases, the system automatically increases the baseline adjustment amount to positively compensate for the initial fusion value, thereby offsetting the systematic shift in pulverized coal characteristics caused by the increased operating intensity of the coal mill under high load conditions. Simultaneously, when the pulverized coal feeding system flow stability index decreases, the system increases the fluctuation compensation factor to exert a stronger fluctuation suppression effect on the initial fusion value, effectively smoothing the instantaneous noise caused by airflow fluctuations or pipeline blockages. This dual adjustment mechanism forms a closed loop with the calculation process of the initial fusion value, enabling the final output pulverized coal quality index analysis results to adapt to equipment load fluctuations and changes in pulverized coal feeding system stability in real time, avoiding the distortion of analysis results caused by dynamic changes in operating conditions in traditional methods.

[0150] As a preferred embodiment, the specific implementation of this application is as follows: In a coal-fired power plant pulverizing system, when the pulverizer current increases and the inlet and outlet pressure difference increases, the system calculates an increase in the equipment load state index, and accordingly determines that the baseline adjustment amount should be increased accordingly. Simultaneously, when the primary air volume fluctuation amplitude expands and the air pressure standard deviation increases, the system calculates a decrease in the flow stability index of the pulverizing system, and accordingly determines that the fluctuation compensation factor should be increased accordingly. The system weights and sums the material characteristic correction parameters and the flow uniformity index according to preset weights to obtain an initial fusion value. Then, it applies the baseline adjustment amount for additive compensation and uses the fluctuation compensation factor to apply multiplicative attenuation processing to the compensated value, finally outputting the corrected pulverized coal quality index analysis results. In this embodiment, the processing unit can specifically be an industrial-grade embedded processor, which collects dynamic parameters of the pulverizing process in real time and executes the above-mentioned adjustment algorithm to ensure that the analysis results accurately reflect the pulverized coal quality status even when operating conditions change.

[0151] Through the above solution, this application solves the problem of distortion in pulverized coal quality analysis results caused by dynamic changes in operating conditions. It ensures that the pulverized coal quality index analysis results can accurately reflect the true quality status under conditions of equipment load fluctuations and changes in the stability of the pulverizing system, thereby improving the robustness and engineering applicability of the analysis results and providing a reliable basis for the optimization of the pulverizing system.

[0152] In practical applications, some of the embodiments described above in this application propose to generate optimized parameters for the pulverizing system based on multi-source decision fusion to improve the accuracy of coal powder quality analysis. However, in its implementation, relying solely on a single optimization objective or simple rules to generate optimized parameters makes it impossible to simultaneously balance the stability of coal powder quality indicators and the energy consumption efficiency of the pulverizing system. Specifically, quality stability optimization may increase energy consumption, or energy consumption optimization may cause quality fluctuations, making it difficult to adapt to the refined control requirements for the coordinated improvement of quality stability and energy efficiency under complex working conditions.

[0153] In response, this application further proposes to input the coal powder quality index analysis results and distribution stability parameters into a multi-objective optimization model, with the highest quality index stability and the lowest unit energy consumption of the pulverizing system as the synergistic optimization objectives, to obtain the optimization parameters of the pulverizing system, including: The first optimization objective is to minimize the standard deviation of the coal powder quality index analysis results, and the second optimization objective is to minimize the calculated value of the unit energy consumption of the pulverizing system. A multi-objective optimization model is constructed with the distribution stability parameter not lower than a preset threshold as a constraint. A multi-objective particle swarm optimization algorithm is used to solve the multi-objective optimization model. Through iterative search by updating particle position and velocity, a set of Pareto optimal solutions that achieve a balance between the first and second optimization objectives is obtained. Based on the preset decision rules, an optimal solution is selected from the Pareto optimal solution set, and the operation parameter vector corresponding to the optimal solution is output as the optimization parameters of the pulverizing system.

[0154] Among them, the coal powder quality index analysis results refer to comprehensive quantitative indicators reflecting the quality status of coal powder, which can be achieved by weighted fusion of multi-source data or output of machine learning models, aiming to provide an objective benchmark for quality assessment; the distribution stability parameter refers to the characteristic quantity characterizing the spatial distribution stability of coal powder flow, which can be achieved by cross-correlation coefficient variance statistics or time delay distribution characteristic analysis, aiming to dynamically quantify the flow stability level; the multi-objective optimization model refers to the mathematical framework for coordinating conflict optimization objectives, which can be constructed using constraint optimization or goal programming methods, aiming to establish a synergistic optimization mechanism between quality stability and energy consumption efficiency; the minimum standard deviation in the first optimization objective refers to minimizing the fluctuation range of quality indicators, which can be achieved based on time series statistical calculations, aiming to control quality stability; the minimum calculated unit energy consumption in the second optimization objective refers to minimizing the energy consumption per unit output of the pulverizing system, which can be calculated based on energy consumption models and real-time operating data, aiming to directly improve system energy efficiency; the distribution stability parameter in the constraint condition is not lower than the preset threshold. The value refers to the set stability lower limit threshold, which can be achieved through dynamic threshold comparison or adaptive boundary adjustment, aiming to prevent the optimization process from sacrificing flow stability; the multi-objective particle swarm optimization algorithm refers to a heuristic search method based on swarm intelligence, which can be implemented using improved particle swarm variants or hybrid evolutionary strategies, aiming to efficiently handle nonlinear objective coupling relationships; particle position update and velocity update iterative search refers to the dynamic adjustment mechanism of particle states in the algorithm, which can be implemented based on iterative formulas of inertia weights and learning factors, aiming to avoid local optimum traps; the Pareto optimal solution set refers to the set of non-dominated solutions, which can be generated using congestion calculation or diversity maintenance strategies, aiming to provide multi-dimensional balance solutions; the preset decision rule refers to the criterion for selecting the optimal solution from the solution set, which can be implemented using weighted scoring or preference function mapping, aiming to adapt to actual production needs; the operating parameter vector refers to the parameter combination controlling the pulverizing system, which can include operating variables such as coal mill speed and air volume adjustment, aiming to directly guide the adjustment of system parameters.

[0155] Specifically, the proposed solution simultaneously inputs the coal powder quality index analysis results and distribution stability parameters into a multi-objective optimization model. First, it uses minimizing the standard deviation of the quality index and minimizing unit energy consumption as dual objective functions, while setting the distribution stability parameters as hard constraints to ensure that the flow stability remains above a safe threshold throughout the optimization process. Based on this, a multi-objective particle swarm optimization algorithm is used to perform iterative search. Through a dynamic update mechanism of particle position and velocity, feasible regions are explored in the solution space, gradually converging to a Pareto optimal solution set. This solution set fully covers the trade-off boundary between quality stability and energy efficiency. Subsequently, the solution set is filtered based on a preset decision rule that comprehensively considers the power plant's operating preferences and real-time operating characteristics, extracting the optimal solution that best meets the current requirements. Finally, this optimal solution is mapped to a specific operating parameter vector and output as the optimization parameters of the pulverizing system to the control system. This process achieves dynamic synergistic optimization of quality stability and energy efficiency, ensures the adaptability of the optimization results to different operating conditions through constraints, and utilizes the multiple-solution characteristic of the Pareto solution set to address the volatility of complex operating conditions.

[0156] As a preferred embodiment, the specific implementation of this application is as follows: In the pulverizing system of a coal-fired power plant, an embedded ARM Cortex-M7 microcontroller is used as the execution unit of the optimization algorithm. The real-time acquired coal powder quality index analysis results and distribution stability parameters are input into a pre-constructed multi-objective optimization model. The model uses the standard deviation of the quality index and unit energy consumption as dual objective functions, and sets the threshold of the distribution stability parameter to 0.85 (determined based on historical data statistics). After starting the multi-objective particle swarm optimization algorithm, the algorithm initializes 50 particles. During the iteration process, the particle motion direction is adjusted according to the velocity update formula. After 200 iterations, it converges to a Pareto optimal solution set containing 15 non-dominated solutions. Subsequently, based on the power plant's "prioritize quality stability" decision rule, the solution with the smallest standard deviation of the quality index and energy consumption lower than the benchmark value is selected from the solution set. Finally, the operation parameter vector corresponding to the solution (including the adjustment amount of the coal mill speed, the set value of the primary air volume, etc.) is output to the pulverizing system actuator to realize the dynamic optimization and adjustment of parameters.

[0157] Through the above-mentioned solution, this application solves the problem of imbalance between quality stability and energy efficiency caused by a single optimization objective. It simultaneously achieves stable control of pulverized coal quality indicators and optimization of energy consumption in the pulverizing system under complex working conditions, avoids the risk of quality fluctuations or abnormal increases in energy consumption, and improves the adaptability and control precision of the pulverizing system in dynamic working conditions.

[0158] All of the above-mentioned optional technical solutions can be combined in any way to form the optional embodiments of this application, and will not be described in detail here.

[0159] Figure 7This is a schematic diagram of the structure of a coal powder quality data analysis system provided in an embodiment of this application. See also... Figure 7 The system includes: The acquisition module 701 is used to acquire multi-source heterogeneous data of coal powder in response to the quality analysis command of coal powder in the target pipeline. The multi-source heterogeneous data of coal powder includes atomic emission spectrum data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data and dynamic parameters of the pulverizing process. The first processing module 702 is used to perform first-type data processing on atomic emission spectrum data and three-dimensional holographic reconstruction data to obtain material property correction parameters and particle characterization features, and generate working condition adaptation parameters based on the dynamic parameters of the powder making process. The second processing module 703 is used to perform second-type data processing on coal flow acoustic data, space charge distribution data and particle characterization characteristics to obtain coal powder flow uniformity index and distribution stability parameters. The generation module 704 is used to generate coal powder quality index analysis results and pulverizing system optimization parameters through multi-source decision fusion based on material property correction parameters, operating condition adaptation parameters, flow uniformity index, and distribution stability parameters. Among them, the weight allocation of each data source in the multi-source decision fusion is dynamically adjusted based on the distribution stability parameters.

[0160] It should be noted that the coal powder quality data analysis system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer equipment can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the coal powder quality data analysis system and the coal powder quality data analysis method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0161] Figure 8 This is a schematic diagram of a server structure provided in an embodiment of this application. The server 800 can vary significantly due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 801 and one or more memories 802. The one or more memories 802 store at least one computer program, which is loaded and executed by the one or more processors 801 to implement the methods provided in the various method embodiments described above. Of course, the server 800 may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server 800 may also include other components for implementing device functions, which will not be elaborated upon here.

[0162] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program that can be executed by a processor to perform the coal powder quality data analysis method in the above embodiments. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage device, etc.

[0163] In an exemplary embodiment, a computer program product or computer program is also provided, which includes program code stored in a computer-readable storage medium. The processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to perform the above-described method for analyzing coal powder quality data.

[0164] In some embodiments, the computer program involved in the present application embodiments may be deployed and executed on a computer device, or executed on multiple computer devices located in one location, or executed on multiple computer devices distributed in multiple locations and interconnected through a communication network. Multiple computer devices distributed in multiple locations and interconnected through a communication network may constitute a blockchain system.

[0165] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0166] The above are merely optional embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for analyzing coal powder quality data, characterized in that, The method includes: In response to the command to analyze the quality of coal powder in the target pipeline, multi-source heterogeneous data of coal powder is acquired. The multi-source heterogeneous data of coal powder includes atomic emission spectroscopy data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data, and dynamic parameters of the pulverizing process. The atomic emission spectral data is subjected to spectral feature analysis and background noise suppression to obtain the denoised elemental characteristic intensity distribution and spectral profile parameters. The three-dimensional holographic reconstruction data is subjected to multi-scale morphology analysis and spatial distribution reconstruction to obtain particle surface morphology features and three-dimensional spatial distribution density. Cross-validation compensation processing is performed based on the elemental characteristic intensity distribution and particle surface morphology features to obtain material property correction parameters. Simultaneously, the chemical activity index of the particles is determined based on the half-width at half-maximum and peak asymmetry in the spectral profile parameters. The flow characteristic index of the particles is determined based on the local density extrema and density gradient distribution in the three-dimensional spatial distribution density. The chemical activity index and the flow characteristic index are weighted and fused to obtain comprehensive particle characterization features. The spatial distribution uniformity of the comprehensive particle characterization features is corrected based on the three-dimensional spatial distribution density to obtain particle characterization features. Finally, operating condition adaptation parameters are generated based on the dynamic parameters of the powder preparation process. The energy distribution characteristics of the coal flow acoustic data in a preset frequency band are extracted, and the energy distribution characteristics are compensated and corrected based on the particle morphology parameters in the particle characterization characteristics to obtain acoustic characteristic parameters. Multi-probe signal correlation analysis is performed on the space charge distribution data to determine the cross-correlation coefficients between different probe signals. These cross-correlation coefficients are then weighted using the surface characteristic parameters in the particle characterization characteristics to obtain electrostatic characteristic parameters. Based on the acoustic and electrostatic characteristic parameters, a consistency index between the acoustic and electrostatic characteristic parameters is determined. When the consistency index exceeds a preset threshold, a coal powder flow uniformity index is generated based on the weighted average of the acoustic and electrostatic characteristic parameters, and a distribution stability parameter is determined based on the variance of the cross-correlation coefficients. Based on the distribution stability parameter, the weights of the material property correction parameter, the operating condition adaptation parameter, and the flow uniformity index in the decision fusion are dynamically allocated to obtain a weight allocation scheme, wherein the distribution stability parameter is positively correlated with the weight of the corresponding data source; based on the weight allocation scheme, the material property correction parameter and the flow uniformity index are weighted and fused, and the weighted fusion process is dynamically adjusted using the operating condition adaptation parameter to obtain the coal powder quality index analysis results; the coal powder quality index analysis results and the distribution stability parameter are jointly input into a multi-objective optimization model, with the highest quality index stability and the lowest unit energy consumption of the pulverizing system as the collaborative optimization objectives, to solve for the pulverizing system optimization parameters.

2. The method according to claim 1, characterized in that, The process of performing spectral feature analysis and background noise suppression on the atomic emission spectrum data to obtain the denoised elemental characteristic intensity distribution and spectral profile parameters includes: The atomic emission spectrum data is decomposed into different frequency bands by wavelet packet transform processing. An adaptive threshold is set using the energy distribution characteristics of each frequency band, and the high-frequency noise components after decomposition are suppressed based on the corresponding adaptive threshold. The denoised spectral signal is then reconstructed based on the suppressed frequency band components. The denoised spectral signal is subjected to continuous spectrum background subtraction, and an adaptive iterative fitting algorithm is used to eliminate background interference caused by matrix effects, so as to obtain pure elemental characteristic spectral lines. The overlapping spectral lines in the pure element feature spectral lines are deconvolved, and a Gaussian mixture model is used to separate the overlapping spectral lines in the overlapping spectral lines into multiple single-element spectral line components. Multi-peak fitting analysis is performed on the multiple single-element spectral line components to obtain the elemental characteristic intensity distribution of each element's characteristic peak and the spectral line profile parameters, wherein the spectral line profile parameters include full width at half maximum (FWHM), peak asymmetry, and spectral line centroid position.

3. The method according to claim 1, characterized in that, The generation of operating condition adaptation parameters based on the dynamic parameters of the powder-making process includes: The operating parameters of the coal mill are extracted from the dynamic parameters of the pulverizing process, and the equipment load status index is determined based on the coal mill current and inlet / outlet pressure difference in the operating parameters of the coal mill. The powder feeding system parameters are extracted from the dynamic parameters of the powder making process, and the flow stability index of the powder feeding system is determined based on the primary air volume and air pressure fluctuation characteristics in the powder feeding system parameters. Based on the equipment load state index and the powder feeding system flow stability index, operating condition evaluation parameters are generated. Based on the operating condition status evaluation parameters, the operating condition adaptation parameters are determined through the operating condition matching model.

4. The method according to claim 1, characterized in that, The step of performing multi-probe signal correlation analysis on the space charge distribution data to determine the cross-correlation coefficients between signals from different probes includes: The charge signals of each probe are divided into multiple time segments by a sliding time window to obtain signal sequences for multiple time periods; Based on the spatial arrangement of multiple probes, determine the probe combinations that need to be correlated and determine the cross-correlation coefficient between the signal sequences of two probes in each probe combination. Determine the peak cross-correlation coefficient and corresponding time delay for each probe combination; Based on the peak cross-correlation coefficients and corresponding time delays of all probe combinations, a probe signal correlation matrix is ​​constructed. The cross-correlation coefficient is determined based on the correlation strength distribution of each probe combination in the probe signal correlation matrix.

5. The method according to claim 1, characterized in that, The coal powder quality index analysis results and the distribution stability parameters are jointly input into a multi-objective optimization model. The model aims to achieve the highest quality index stability and the lowest unit energy consumption of the pulverizing system as the synergistic optimization objectives, and solves for the optimized parameters of the pulverizing system, including: The first optimization objective is to minimize the standard deviation of the coal powder quality index analysis results, the second optimization objective is to minimize the calculated value of the unit energy consumption of the pulverizing system, and the constraint condition is that the distribution stability parameter is not lower than a preset threshold. The multi-objective optimization model is constructed accordingly. The multi-objective particle swarm optimization algorithm is used to solve the multi-objective optimization model. Through iterative search by updating particle position and velocity, a set of Pareto optimal solutions that achieve a balance between the first optimization objective and the second optimization objective is obtained. Based on preset decision rules, an optimal solution is selected from the Pareto optimal solution set, and the operation parameter vector corresponding to the optimal solution is output as the optimization parameters of the pulverizing system.

6. A system for analyzing pulverized coal quality data, characterized in that, The system includes: The acquisition module is used to acquire multi-source heterogeneous data of coal powder in response to the quality analysis command of coal powder in the target pipeline. The multi-source heterogeneous data of coal powder includes atomic emission spectrum data, three-dimensional holographic reconstruction data, coal flow acoustic data, space charge distribution data and dynamic parameters of the pulverizing process. The first processing module is used to perform spectral feature analysis and background noise suppression on the atomic emission spectral data to obtain the denoised elemental characteristic intensity distribution and spectral profile parameters; to perform multi-scale morphology analysis and spatial distribution reconstruction on the three-dimensional holographic reconstruction data to obtain particle surface morphology features and three-dimensional spatial distribution density; to perform cross-validation compensation processing based on the elemental characteristic intensity distribution and particle surface morphology features to obtain material property correction parameters; to determine the chemical activity index of the particles based on the half-width at half-maximum and peak asymmetry in the spectral profile parameters; to determine the flow characteristic index of the particles based on the local density extrema and density gradient distribution in the three-dimensional spatial distribution density; to perform weighted fusion of the chemical activity index and the flow characteristic index to obtain comprehensive particle characterization features; to perform spatial distribution uniformity correction on the comprehensive particle characterization features based on the three-dimensional spatial distribution density to obtain particle characterization features; and to generate operating condition adaptation parameters based on the dynamic parameters of the powder making process. The second processing module is used to extract the energy distribution characteristics of the coal flow acoustic data in a preset frequency band, and simultaneously compensate and correct the energy distribution characteristics based on the particle morphology parameters in the particle characterization characteristics to obtain acoustic characteristic parameters; perform multi-probe signal correlation analysis on the space charge distribution data to determine the cross-correlation coefficients between different probe signals, and weight the cross-correlation coefficients in combination with the surface characteristic parameters in the particle characterization characteristics to obtain electrostatic characteristic parameters; determine the consistency index between the acoustic characteristic parameters and the electrostatic characteristic parameters based on the acoustic characteristic parameters and the electrostatic characteristic parameters; when the consistency index exceeds a preset threshold, generate a coal powder flow uniformity index based on the weighted average of the acoustic characteristic parameters and the electrostatic characteristic parameters, and determine the distribution stability parameter based on the variance of the cross-correlation coefficients; A generation module is used to dynamically allocate the weights of the material property correction parameter, the operating condition adaptation parameter, and the flow uniformity index in the decision fusion based on the distribution stability parameter, thereby obtaining a weight allocation scheme. The distribution stability parameter is positively correlated with the weight of the corresponding data source. Based on the weight allocation scheme, the material property correction parameter and the flow uniformity index are weighted and fused, and the weighted fusion process is dynamically adjusted using the operating condition adaptation parameter to obtain the coal powder quality index analysis results. The coal powder quality index analysis results and the distribution stability parameter are jointly input into a multi-objective optimization model, with the highest quality index stability and the lowest unit energy consumption of the pulverizing system as the collaborative optimization objectives, to solve for the optimized parameters of the pulverizing system.

Images