Intelligent analysis method and system for coal quality detection data

Abstract

The invention relates to the technical field of coal-fired power generation, and discloses a coal quality detection data intelligent analysis method and system.The method comprises the steps that cyclic sampling operation is executed at multiple spatial positions of an outlet pipeline of a coal mill, and a pulverized coal sample is obtained; quantitatively intercepting the pulverized coal sample, and applying preset pressure to form a to-be-measured entity; performing multi-physical quantity signal acquisition on an entity to be detected to obtain a dielectric characteristic signal reflecting the moisture content, a molecular vibration spectrum signal reflecting organic components and an atomic characteristic spectrum signal reflecting mineral elements; extracting characteristic values in the dielectric characteristic signal, the molecular vibration spectrum signal and the atomic characteristic spectrum signal, substituting the characteristic values into a preset quantitative correlation equation set, synchronously solving to obtain a moisture index, an ash index and a calorific value index, comparing the moisture index, the ash index and the calorific value index with corresponding preset reference values, generating and executing a coal mill adjusting instruction, and performing fusion analysis on multiple physical quantity signals to obtain a coal mill adjusting result. And an accurate data basis is provided for real-time fine adjustment of boiler combustion.

Description

Technical Field

[0001] This application relates to the field of coal-fired power generation technology, and more specifically, to an intelligent analysis method and system for coal quality testing data. Background Technology

[0002] As a crucial component of energy supply, the economic and environmental performance of coal-fired power generation is closely linked to the stability of the coal quality fed into the furnace. With increasingly stringent grid requirements for peak-shaving capacity, generating units need to frequently perform deep peak-shaving and low-load flexible operation. Under these conditions, even minor fluctuations in coal quality can significantly impact the combustion stability and safety of the boiler. However, the industry currently relies heavily on traditional offline laboratory testing for coal quality analysis. This process typically involves manual sampling, sample delivery, sample preparation, and lengthy chemical analysis, with the entire cycle potentially lasting several hours or even longer. This significant data lag poses a critical technical problem: by the time the test results are finally available, the batch of coal it represents has already been completely burned, rendering the results historical data and unable to reflect the true quality of the pulverized coal currently being fed into the mill and boiler. Consequently, the boiler's combustion control system cannot perform feedforward adjustments based on real-time, accurate coal quality parameters, and can only perform delayed feedback control. This leads to low combustion efficiency, increased pollutant emissions, and a high risk of unstable combustion or even flameout during rapid load changes.

[0003] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides an intelligent analysis method and system for coal quality testing data, which can solve the technical problems of data lag and inability to provide real-time and effective data support for boiler combustion in traditional coal quality testing methods.

[0005] Firstly, this application provides an intelligent analysis method for coal quality testing data, including: Cyclic sampling was performed at multiple spatial locations in the coal mill outlet pipe to obtain pulverized coal samples; A quantitative sample of pulverized coal is extracted and a preset pressure is applied to form the sample to be tested. Multi-physical quantity signal acquisition is performed on the entity to be tested to obtain dielectric property signals reflecting moisture content, molecular vibrational spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. The characteristic values ​​of dielectric properties, molecular vibrational spectrum and atomic characteristic spectrum are extracted and substituted into the preset quantitative correlation equations to simultaneously solve for moisture, ash and calorific value. The obtained moisture, ash, and calorific value indices are compared with the corresponding preset benchmark values ​​to generate and execute the coal mill adjustment command.

[0006] By constructing a closed-loop online analysis process from sampling, sample preparation, testing to solution and control, and through the fusion analysis of multiple physical quantity signals, it is possible to quickly and synchronously obtain multiple key coal quality indicators such as moisture, ash content, and calorific value. This overcomes the shortcomings of traditional testing methods, such as long time consumption and data lag, and provides an accurate data foundation for the real-time and refined adjustment of boiler combustion.

[0007] Furthermore, the steps of performing cyclic sampling at multiple spatial locations at the coal mill outlet to obtain pulverized coal samples include: Sampling channels distributed at multiple spatial locations in the coal mill outlet pipeline are opened in turn according to preset time intervals. A negative pressure environment is established in the sampling channel to guide the coal powder airflow in the coal mill outlet pipe to the separation path, and the physical separation of coal powder particles in the coal powder airflow from the carrier gas is achieved by centrifugal force to obtain a coal powder sample.

[0008] By adopting a multi-point cyclic sampling method, the representativeness of single-point sampling, which may be insufficient due to uneven distribution of coal powder in the pipeline, is effectively overcome. At the same time, by using negative pressure drainage and centrifugal separation, fully automatic and efficient collection of coal powder samples is achieved, laying a solid foundation for subsequent accurate analysis.

[0009] Furthermore, the steps of establishing a negative pressure environment within the sampling channel to guide the pulverized coal airflow in the coal mill outlet pipe to the separation path, and using centrifugal force to physically separate the pulverized coal particles from the carrier gas to obtain a pulverized coal sample include: The coal powder airflow is guided to the primary separation zone, and the first centrifugal force field is used to capture the first part of the coal powder particles and make them fall into the collection zone. The coal powder airflow after passing through the primary separation zone is guided to the secondary separation zone for secondary separation. The second centrifugal force field is used to capture fine dust and make it fall into the collection area. The coal powder particles and fine dust in the first part of the collection area are combined to obtain a coal powder sample.

[0010] By adopting a two-stage centrifugal separation design, coal powder particles of different sizes can be captured in stages and efficiently. In particular, it can effectively recover fine dust that is difficult to capture by traditional single-stage separators, thereby ensuring the integrity and representativeness of the final coal powder sample and improving the accuracy of the test results.

[0011] Furthermore, the steps of quantitatively extracting the coal powder sample and applying a preset pressure to form the test substance include: A sample of coal powder of a predetermined weight is taken, a predetermined pressure is applied and maintained for a predetermined time to obtain a test entity with a predetermined saturation density.

[0012] By standardizing the quantification and density of coal powder samples, the interference caused by the inconsistency between sample weight and compaction density on the subsequent acquisition of physical quantity signals is effectively eliminated, ensuring that each measurement is carried out under the same physical benchmark, and significantly improving the stability and comparability of the test results.

[0013] Furthermore, the steps of applying a preset pressure and maintaining it for a preset duration include: Real-time monitoring of the correlation curve between pressure intensity and displacement change during the pressure application process; The elastic characteristics of the pulverized coal sample were determined based on the correlation curve. When the elastic characteristics meet the preset rebound conditions, the preset time is extended to eliminate internal stress.

[0014] By monitoring the pressure-displacement curve during the pressing process in real time, the elastic behavior of pulverized coal can be dynamically determined, and the holding time can be adaptively adjusted to release internal stress, preventing the test object from cracking or having uneven density due to rebound, thus further ensuring the high quality of the test object and the accuracy of subsequent measurements.

[0015] Furthermore, prior to the step of acquiring multi-physical quantity signals from the entity under test, the following steps are included: Perform a topography scan on the surface of the entity to be tested to obtain surface integrity features; When the surface integrity feature is lower than a preset threshold, a morphology correction coefficient is generated based on the distribution density of surface defects in the entity under test. After acquiring the dielectric property signal reflecting moisture content, the molecular vibrational spectral signal reflecting organic components, and the atomic characteristic spectral signal reflecting mineral elements, the process includes: The morphology correction coefficient is used to numerically compensate for the collected molecular vibrational spectral signals in order to offset the diffuse reflection loss.

[0016] By adding an evaluation step of the surface morphology of the target object before formal signal acquisition, and generating correction coefficients based on surface defects, the loss of diffuse reflection light caused by the unevenness of the sample surface is actively compensated, the intensity of the spectral signal is effectively corrected, the interference of physical morphology on chemical composition analysis is eliminated, and the accuracy of spectral analysis is improved.

[0017] Furthermore, the steps for acquiring multiple physical quantity signals of the entity under test within the same processing cycle include: Obtain the initial surface temperature, ambient temperature, and ambient humidity of the object under test; The water evaporation intensity is calculated based on the difference between the initial surface temperature and the ambient temperature, as well as the ambient humidity. Based on the water evaporation intensity, a preset attenuation compensation function is invoked to increase the water response weight in the dielectric characteristic signal in real time, thereby obtaining the compensated dielectric characteristic signal.

[0018] By calculating the evaporation intensity in real time and dynamically adjusting the response weight of the dielectric signal, the signal attenuation caused by natural moisture loss during the measurement process can be accurately compensated, ensuring the immediacy and accuracy of moisture index measurement and avoiding measurement errors introduced by environmental changes.

[0019] Furthermore, the steps of extracting eigenvalues ​​from dielectric property signals, molecular vibrational spectral signals, and atomic characteristic spectral signals, and substituting them into a pre-set quantitative correlation equation set to simultaneously solve for moisture content, ash content, and calorific value include: The resonant frequency offset and quality factor are extracted from the dielectric property signal as dielectric characteristic values. The absorbance or reflectance at a preset wavelength position is extracted from the molecular vibrational spectral signal as a molecular vibrational characteristic value. The characteristic peak intensities of preset mineral elements are extracted from the atomic characteristic spectrum signal and used as the characteristic values ​​of the atomic characteristic spectrum. Based on the characteristic peak intensity of preset mineral elements, the coal type classification information of the current coal powder sample is identified; Retrieve a set of target quantitative correlation equations that match the coal type classification information from a pre-defined equation database; The extracted resonant frequency shift, quality factor, absorbance or reflectance, and characteristic peak intensity are substituted into the target quantitative correlation equations for solution to obtain moisture index, ash index, and calorific value index.

[0020] This technical solution not only clarifies the key feature values ​​extracted from multi-source signals, but also innovatively utilizes atomic feature spectrum signals to achieve online identification of coal types. Based on the identification results, it retrieves a matching, dedicated quantitative analysis model from the database. This adaptive modeling strategy greatly improves the model's universality for different coal types and the accuracy of the analysis results, solving the problem that traditional fixed models struggle to adapt to changes in coal sources.

[0021] Furthermore, the steps of comparing the obtained moisture, ash, and calorific value indices with the corresponding preset benchmark values ​​to generate and execute the coal mill adjustment command include: When the moisture index is higher than the preset moisture benchmark value, the adjustment command to reduce the coal feed rate of the coal mill and increase the drying air volume is executed. When the ash content index is higher than the preset ash content benchmark value, the adjustment command for blowing soot onto the boiler heating surface is executed. When the calorific value is lower than the preset calorific value benchmark, an adjustment command is executed to increase the output of the coal mill or increase the primary air temperature.

[0022] This technical solution compares the analyzed coal quality indicators with preset benchmark values ​​and generates clear and specific adjustment instructions for the coal mill and boiler accordingly. This achieves seamless integration from online detection to closed-loop control, making combustion adjustment no longer dependent on manual experience, but based on real-time data for automated and precise feedforward control, thereby improving the response speed and operating efficiency of the entire combustion system.

[0023] Secondly, this application also discloses an intelligent analysis system for coal quality testing data, used to perform the intelligent analysis method for coal quality testing data described in any of the foregoing claims, comprising: The sampling module is used to perform cyclic sampling operations at multiple spatial locations in the coal mill outlet pipe to obtain coal powder samples; The sample preparation module is used to quantitatively extract coal powder samples and apply a preset pressure to form the sample to be tested. The signal acquisition module is used to perform multi-physical quantity signal acquisition on the entity under test, and obtain dielectric property signals reflecting moisture content, molecular vibration spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. The index solving module is used to extract the characteristic values ​​from the dielectric property signal, molecular vibrational spectrum signal and atomic characteristic spectrum signal, and substitute them into the preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index and calorific value index. The control execution module is used to compare the obtained moisture, ash and calorific value indicators with the corresponding preset benchmark values, and generate and execute the coal mill adjustment command.

[0024] This application provides an intelligent analysis method and system for coal quality testing data, which, compared to existing technologies, has at least the following advantages: First, by performing online sampling, sample preparation, and testing at the coal mill outlet, this application shortens the traditional offline testing process, which can take several hours, to minutes, completely solving the problem of severe data lag in coal quality analysis and ensuring the real-time nature of the test results. Second, this application innovatively integrates multiple physical detection methods such as dielectric, spectral, and atomic spectroscopy, enabling the simultaneous and rapid acquisition of multiple interrelated key indicators such as moisture, ash content, and calorific value. Compared to traditional single-indicator detection, the analysis results are more accurate and reliable. Finally, this application directly uses the real-time analysis results to generate and execute adjustment commands for the coal mill and boiler, constructing a fully automated system from real-time detection to intelligent analysis and closed-loop control. This achieves feedforward adjustment based on coal quality changes, effectively improving the stability and economy of boiler combustion, reducing pollutant emissions, and is particularly adaptable to the stringent requirements of flexible operation such as deep peak shaving in power plants. Attached Figure Description

[0025] Figure 1This is a flowchart illustrating an intelligent analysis method for coal quality testing data provided in an embodiment of this application.

[0026] Figure 2 This is a schematic diagram of the structure of an intelligent analysis system for coal quality testing data provided in an embodiment of this application.

[0027] Labeling Explanation: 210, Sampling Module; 220, Sample Preparation Module; 230, Signal Acquisition Module; 240, Index Solving Module; 250, Control Execution Module. Detailed Implementation

[0028] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0029] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] In the daily operation of coal-fired power plants, especially under deep peak-shaving conditions to cope with rapid fluctuations in grid load, the stability of boiler combustion faces severe challenges. One of the core limiting factors is the significant delay in recognizing the quality of pulverized coal entering the furnace. Traditional coal quality analysis relies on manual sampling and sending samples to the laboratory, a process that takes several hours. By the time the analysis results are available, the corresponding coal has already been consumed. This prevents the combustion control system from making proactive adjustments based on actual coal quality changes, forcing it to react passively. Consequently, this leads to reduced combustion efficiency, increased pollutant emissions, and even the risk of flameout at low loads.

[0031] Regarding this, firstly, see... Figure 1 This application provides an intelligent analysis method for coal quality testing data, including: S1. Perform cyclic sampling at multiple spatial locations in the coal mill outlet pipe to obtain coal powder samples; S2. Quantitatively extract the coal powder sample and apply a preset pressure to form the test object; S3. Perform multi-physical quantity signal acquisition on the entity to be tested to obtain dielectric property signals reflecting moisture content, molecular vibration spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. S4. Extract the characteristic values ​​from the dielectric property signal, molecular vibrational spectrum signal and atomic characteristic spectrum signal, and substitute them into the preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index and calorific value index. S5. Compare the obtained moisture, ash and calorific value indices with the corresponding preset benchmark values, and generate and execute the coal mill adjustment command.

[0032] The test object refers to a loose coal powder sample that has been compressed into a solid block with a specific geometry, size, and density through a standardized preparation process. The reason for forming a test object is that the physical morphology of loose powder is extremely unstable; its bulk density and interparticle spacing change with even minor perturbations. This can cause significant interference to subsequent optical, electromagnetic, and other physical signal acquisition processes, leading to poor repeatability of measurement results. By preparing it into a standardized test object, a stable and consistent measurement target can be provided for various physical detection methods, which is a prerequisite for achieving multi-signal fusion analysis and ensuring measurement accuracy.

[0033] Multi-physical quantity signal acquisition refers to the simultaneous or quasi-simultaneous acquisition of a series of signals characterizing the different physicochemical properties of the same analyte using multiple sensors or detection instruments based on different physical principles within the same processing cycle. For example, moisture can be detected using the principle of electromagnetic field-matter interaction, organic components can be detected using molecular vibrational spectroscopy, and inorganic mineral elements can be detected using atomic emission spectroscopy. The core of this method lies in detecting the sample from multiple dimensions, obtaining more comprehensive and accurate information on the material composition than a single detection method through information complementarity and cross-validation.

[0034] Quantitative correlation equations refer to a set of mathematical models used to describe the complex nonlinear relationships between multiple input physical quantities and multiple output chemical indicators. These equations are typically established based on extensive experimental data, involving the acquisition of multiple physical quantity signals from hundreds or thousands of coal samples with known chemical compositions (accurately determined using traditional testing methods). Chemometric algorithms, such as multiple linear regression, partial least squares, or artificial neural networks, are then used to uncover the mathematical patterns between the physical signal characteristic values ​​and indicators such as moisture, ash content, and calorific value. Once the equation set is established, for unknown samples, their chemical indicators can be quickly deduced simply by substituting their physical signal characteristic values.

[0035] In one specific embodiment, the system first needs to acquire representative samples from the high-speed conveying of pulverized coal gas flow. The diameter of the outlet pipe of the coal mill is typically large, and the concentration and particle size distribution of the pulverized coal particles inside are uneven across the pipe cross-section due to factors such as gravity and airflow turbulence. To overcome the limitations of single-point sampling, the system installs sampling probes at different locations along the pipe, such as top, bottom, left, and right. The central controller, according to a preset program, for example, opens the sampling valve of one of the probes in turn every five minutes for brief sampling, thereby achieving cyclic scanning sampling of the entire pipe cross-section and ensuring the overall representativeness of the collected samples.

[0036] After obtaining the coal powder sample, it is transferred to the sample preparation unit. Here, the system first uses a precision weighing mechanism to accurately extract a predetermined weight of coal powder, for example, five grams, from the collected sample. This is done to eliminate systematic errors caused by inconsistent sample quantities in subsequent analysis. This quantitative amount of coal powder is then fed into a compaction mold. A hydraulically or electrically servo-driven pressure head applies a predetermined pressure, for example, 20 MPa, to the coal powder within the mold and maintains it for a predetermined time, for example, ten seconds. Through this compaction process, the loose coal powder is compressed into a disc-shaped or cylindrical sample. This standardized sample preparation process ensures that each sample has a basically consistent weight, shape, and compaction density, providing a stable and reliable measurement object for subsequent signal acquisition.

[0037] Next, the prepared test sample is automatically transferred to the signal acquisition station by a robotic arm. Here, multiple detection probes work in concert. One dielectric sensor probe, which can be shaped like two parallel metal plates, sandwiches the test sample. The sensor emits a high-frequency alternating electric field through the test sample. Water molecules in coal powder are polar molecules; they undergo orientation polarization in the alternating electric field. This process consumes electric field energy and changes the equivalent dielectric constant of the medium. By measuring the frequency shift and quality factor changes of the resonant circuit, the moisture content in the test sample can be accurately determined. Simultaneously, a near-infrared spectroscopy probe emits a broad-spectrum near-infrared light from above the surface of the test sample. After the light shines on the sample surface, part of it is absorbed and part is reflected. Organic matter in coal powder, such as hydrocarbons, has chemical bonds within its molecules that vibrate or rotate under the excitation of infrared light at specific wavelengths, thus absorbing energy at the corresponding wavelength. The spectrometer receives and analyzes the reflected light; by measuring the absorbance at specific wavelengths, the molecular vibrational spectral signals related to the organic components and calorific value of the coal powder can be obtained. In addition, an atomic characteristic spectrum acquisition device, such as a laser-induced breakdown spectrometer, is also aimed at the object under test. The instrument emits a high-energy pulsed laser beam, which ablates a tiny area on the surface of the object under test, instantly generating a high-temperature plasma. During the de-excitation process, the atoms and ions in the plasma radiate characteristic spectral lines representing their respective elemental identities. The spectrometer captures these spectral lines, and by analyzing the characteristic peak intensities of elements such as silicon, aluminum, and iron, atomic characteristic spectral signals related to the mineral content in the coal powder, i.e., the ash content, can be obtained.

[0038] After being digitized by the acquisition system, the three physical signals are transmitted to the data processing unit. The core of this unit is a pre-constructed set of quantitative correlation equations. The processing unit's software first extracts key feature values ​​from the raw signals. For example, it extracts the resonant frequency shift from the dielectric signal, the absorbance values ​​at specific wavelengths from the near-infrared spectrum, and the spectral line intensities representing the main ash components such as silicon dioxide and aluminum oxide from the atomic characteristic spectrum. These feature values ​​are then used as input variables and substituted into the quantitative correlation equations. For example, a typical set of equations might contain equations of the following form: moisture content equals coefficient A multiplied by the frequency shift plus a constant term; ash content equals coefficient B multiplied by the silicon spectral line intensity plus coefficient C multiplied by the aluminum spectral line intensity; and the calorific value might be a more complex function incorporating moisture, ash, and near-infrared spectral feature values. By simultaneously solving this set of equations, the system can calculate the moisture content, ash content, and calorific value of the current coal powder sample within seconds.

[0039] Finally, the calculated real-time coal quality indicators are sent to the control execution unit. This unit compares these real-time values ​​with preset benchmark values ​​stored in the system. These benchmark values ​​are ideal coal quality parameter ranges set according to the current boiler operating conditions and operational goals (such as optimal efficiency or minimum pollutant levels). For example, if the calculated moisture content is 12%, while the upper limit of the benchmark value is 10%, the system will determine that the current pulverized coal moisture content is too high. Accordingly, the system will automatically generate an adjustment command and send it to the power plant's distributed control system via the communication interface. The command will increase the primary air temperature by 5% or decrease the coal mill feed rate by 3% to enhance the drying of the pulverized coal. Similarly, if the ash content exceeds the standard, it may trigger a command to blow soot onto the boiler heating surfaces; if the calorific value is too low, it may trigger a command to increase the coal mill output or co-fire some high-quality coal. Through this series of automated steps, the method of this application constructs a complete link from real-time detection to closed-loop control, enabling the boiler combustion system to anticipate coal quality fluctuations, thereby achieving more stable and efficient operation.

[0040] Furthermore, the steps of performing cyclic sampling at multiple spatial locations at the coal mill outlet to obtain pulverized coal samples include: Sampling channels distributed at multiple spatial locations in the coal mill outlet pipeline are opened in turn according to preset time intervals. A negative pressure environment is established in the sampling channel to guide the coal powder airflow in the coal mill outlet pipe to the separation path, and the physical separation of coal powder particles in the coal powder airflow from the carrier gas is achieved by centrifugal force to obtain a coal powder sample.

[0041] The basic solution described above only mentions opening valves at multiple locations for sampling, but it doesn't detail how to effectively capture coal dust particles from the high-speed airflow. In actual operation, the coal dust airflow velocity at the mill outlet is extremely high. Simply opening a branch channel may not be sufficient to reliably extract a sufficient amount of sample due to pressure balance or disturbance issues. Therefore, this embodiment proposes an improved solution using active suction. Specifically, each sampling channel is connected to a blower or vacuum pump via a pipeline at its rear end. When the central controller decides to sample at a certain location, it not only opens the sampling valve at that location but also simultaneously starts the blower, establishing a stable negative pressure environment throughout the sampling branch. This negative pressure acts like a vacuum cleaner, effectively drawing a portion of the coal dust airflow from the main channel into the sampling channel. This active suction method significantly improves sampling efficiency and stability compared to passive opening. The sucked-in coal dust airflow does not directly enter the collector but is guided to a separation device. A simple and efficient separation device is a cyclone separator. The pulverized coal gas flows tangentially into the cylindrical section of the cyclone separator, creating a high-speed rotating vortex. Under the powerful centrifugal force, pulverized coal particles, much denser than the carrier gas, are thrown against the inner wall of the separator. Upon collision with the inner wall, the particles lose kinetic energy and spiral down along the wall under gravity, eventually falling into the collection hopper at the bottom. The purified carrier gas is discharged from the central outlet at the top of the separator. This combination of negative pressure flow and centrifugal separation enables fully automated, efficient, and continuous collection of pulverized coal samples.

[0042] Furthermore, the steps of establishing a negative pressure environment within the sampling channel to guide the pulverized coal airflow in the coal mill outlet pipe to the separation path, and using centrifugal force to physically separate the pulverized coal particles from the carrier gas to obtain a pulverized coal sample include: The coal powder airflow is guided to the primary separation zone, and the first centrifugal force field is used to capture the first part of the coal powder particles and make them fall into the collection zone. The coal powder airflow after passing through the primary separation zone is guided to the secondary separation zone for secondary separation. The second centrifugal force field is used to capture fine dust and make it fall into the collection area. The coal powder particles and fine dust in the first part of the collection area are combined to obtain a coal powder sample.

[0043] While the aforementioned single-stage cyclone separation scheme is feasible, it has an inherent limitation: its separation efficiency varies depending on the particle size. For larger, heavier coal dust particles, a single-stage cyclone separator can effectively capture them. However, for very fine dust particles (e.g., less than ten micrometers), due to their low inertia, they easily escape from the top outlet following the central streamline of the rotating airflow, resulting in the loss of these fine particles. These fine particles also play an important role in the overall composition and combustion characteristics of coal dust; their absence affects the representativeness of the sample and, consequently, the accuracy of the analytical results. To address this issue, this embodiment proposes a two-stage separation scheme. Specifically, the coal dust airflow drawn from the sampling channel first enters a primary separation zone, typically a conventional cyclone separator designed to handle high concentrations of larger particles. Here, a relatively mild first centrifugal force field is used to capture the vast majority (e.g., over 95%) of the coal dust particles in the airflow, especially the larger particles, and cause them to fall into the first collection zone below. Although the gas exiting from the top of the primary separator appears much cleaner, it still carries a significant amount of fine dust. This gas flow is not directly discharged but is guided to a secondary separation zone. The secondary separation zone can be one or more parallel high-efficiency small-diameter cyclone separators, or a specially designed long conical separator. These secondary separators are structurally designed to generate a stronger second centrifugal force field. When the gas flow carrying fine dust enters the secondary separation zone, the fine dust is effectively thrown against the wall and captured by the stronger centrifugal force, falling into the second collection zone below. Finally, the system combines the large coal particles captured in the first collection zone with the fine dust captured in the second collection zone to form the final coal sample used for analysis. This significantly improves the overall capture efficiency for coal particles across the entire particle size range, ensuring the integrity and representativeness of the obtained sample.

[0044] Furthermore, the steps of quantitatively extracting the coal powder sample and applying a preset pressure to form the test substance include: A sample of coal powder of a predetermined weight is taken, a predetermined pressure is applied and maintained for a predetermined time to obtain a test entity with a predetermined saturation density.

[0045] The basic scheme mentions compressing coal powder, but the objectives and standards for this compression are not clearly defined. Simply specifying the pressure and time does not guarantee that the final test sample will have consistent physical properties. This is because the compressibility of coal powder varies depending on the type of coal and its moisture content. Under the same pressure and time, one type of coal powder may be compressed very densely, while another may remain loose. This inconsistency in density directly affects subsequent measurements of dielectric constant and spectral reflectance, introducing errors that are difficult to eliminate. Therefore, this embodiment clarifies the core objective of sample preparation: obtaining a test sample with a preset saturated density. Saturated density refers to a relatively stable density value at which, under certain compression conditions, the powder material reaches densification, and its density no longer increases significantly with increasing pressure.

[0046] To achieve this goal, the sample preparation process is refined. First, a predetermined weight of coal powder sample is collected using a high-precision electronic balance – this is the first step in density control. Then, during the pressing process, the system not only controls the applied pressure and holding time, but more importantly, these two parameters (pressure and duration) are pre-calibrated experimentally. The calibration process involves pressing various representative coal samples under different pressures and measuring their final density, plotting a pressure-density curve. An inflection point can be found on the curve; beyond this pressure, the increase in density becomes very slow. The system selects this inflection point pressure or slightly higher as the preset pressure, and a sufficient time to allow the internal stress of the sample to fully relax as the preset duration. Through such strict control of weight, pressure, and duration, it can be ensured that regardless of the original state of the coal powder, the final sample will achieve a highly consistent, predetermined density value close to its saturation density, for example, 1.2 grams per cubic centimeter.

[0047] Furthermore, the steps of applying a preset pressure and maintaining it for a preset duration include: Real-time monitoring of the correlation curve between pressure intensity and displacement change during the pressure application process; The elastic characteristics of the pulverized coal sample were determined based on the correlation curve. When the elastic characteristics meet the preset rebound conditions, the preset time is extended to eliminate internal stress.

[0048] The aforementioned scheme, employing fixed preset pressure and duration, while significantly improving density consistency, remains an open-loop control method, assuming that the behavior of all pulverized coal conforms to the average model used during calibration. However, in reality, certain coal types or specific batches of pulverized coal may exhibit strong elasticity. This means that after pressure is released, the sample may rebound to some extent like a spring, resulting in a final density lower than expected, and may even lead to micro-cracks due to uneven internal stress, affecting its structural integrity. This elastic behavior cannot be anticipated or compensated for by a fixed-duration pressing procedure.

[0049] In practice, a high-precision pressure sensor is installed on the pressure head of the pressing device, along with a displacement sensor to measure the pressure head's stroke. Throughout the pressing process, the control system simultaneously records the pressure and displacement values ​​at a high frequency (e.g., 100 times per second) and plots the correlation curve of pressure-displacement changes in real time. The shape of this curve dynamically reflects the mechanical behavior of the pulverized coal. For example, during the holding pressure stage, if the displacement reading quickly stabilizes at a fixed value, it indicates that the sample has good plasticity and the internal stress is quickly released. Conversely, if the displacement still changes slowly during the holding pressure stage (i.e., creep), or if there is a significant reverse change in displacement during the initial stage of depressurization (i.e., rebound), the system can determine that the current pulverized coal sample has strong elastic characteristics based on the characteristics of these correlation curves. The system has a preset rebound condition; for example, this condition is met when the detected displacement rebound exceeds a preset threshold (e.g., 0.05 mm) during the initial stage of depressurization, or when the creep rate during the holding pressure stage is higher than a certain standard. Once the current sample is determined to meet the rebound condition, the control system will automatically extend the preset holding time during the next or current holding phase, for example, from the standard ten seconds to twenty seconds. The purpose of extending the time is to give the sample more time to release its internal elastic stress. This intelligent control method, which adaptively adjusts the holding time, effectively compensates for the elastic differences between different coal samples, maximally suppresses the rebound effect, and ensures that every tested entity, regardless of its characteristics, ultimately reaches a stable, defect-free state that meets the preset density.

[0050] In a preferred embodiment, the step of performing multi-physical quantity signal acquisition on the entity under test includes: Perform a topography scan on the surface of the entity to be tested to obtain surface integrity features; When the surface integrity feature is lower than a preset threshold, a morphology correction coefficient is generated based on the distribution density of surface defects in the entity under test. After acquiring the dielectric property signal reflecting moisture content, the molecular vibrational spectral signal reflecting organic components, and the atomic characteristic spectral signal reflecting mineral elements, the process includes: The morphology correction coefficient is used to numerically compensate for the collected molecular vibrational spectral signals in order to offset the diffuse reflection loss.

[0051] In addition to the aforementioned sensors, the signal acquisition station also integrates a high-resolution surface topography scanning device, such as a line laser profiler or a 3D camera based on structured light technology. Before spectral acquisition, the scanner rapidly scans the upper surface of the object under test, generating a high-precision three-dimensional height map containing millions of data points.

[0052] The data processing unit then analyzes this height map and calculates one or more surface integrity characteristic indicators for evaluating surface quality. A simple indicator could be the standard deviation of surface height, while a more complex indicator could be the defect area ratio, which involves identifying areas with height anomalies (such as pits, bumps, and cracks) using image processing algorithms and calculating the percentage of their area relative to the total area. The system has a preset threshold; for example, when the defect area ratio exceeds three percent, the surface integrity of the entity under test is considered to be below the preset threshold, requiring compensation.

[0053] At this point, the system will invoke a preset function to generate a morphology correction coefficient based on the distribution density and severity of the defects. The basic principle is that the rougher the surface and the more defects there are, the more severe the diffuse reflection of light, resulting in a lower effective signal intensity received by the spectrometer. Therefore, the morphology correction coefficient is usually a value greater than 1, and its magnitude is positively correlated with the defect density. For example, the correction coefficient can be defined as 1 + k*(defect area ratio), where k is an empirical constant obtained through experimental calibration.

[0054] After completing all signal acquisition, before solving for the indicators, the system uses the generated morphology correction coefficient to numerically compensate the original molecular vibrational spectral signal (i.e., near-infrared spectral data). Specifically, the absorbance or reflectance value at each wavelength point on the spectral curve is multiplied by this correction coefficient. This step effectively compensates for signal attenuation caused by surface defects, essentially restoring a flawed sample to an ideal, smooth one at the data level. This eliminates the interference of physical morphology on chemical composition analysis and significantly improves the analytical accuracy of indicators such as calorific value.

[0055] Furthermore, the steps for acquiring multiple physical quantity signals of the entity under test within the same processing cycle include: Obtain the initial surface temperature, ambient temperature, and ambient humidity of the object under test; The water evaporation intensity is calculated based on the difference between the initial surface temperature and the ambient temperature, as well as the ambient humidity. Based on the water evaporation intensity, a preset attenuation compensation function is invoked to increase the water response weight in the dielectric characteristic signal in real time, thereby obtaining the compensated dielectric characteristic signal.

[0056] When the object under test is first removed from the coal mill pipeline and pressed, its temperature may be slightly higher than the surrounding environment. Upon transfer to the signal acquisition station, if the ambient humidity is low, surface moisture will begin to evaporate, resulting in a slightly lower moisture content than its true state within the pipeline during dielectric measurements. To compensate for this loss, the signal acquisition station is equipped with a non-contact infrared thermometer to accurately measure the initial surface temperature of the object under test immediately upon placement. Simultaneously, environmental sensors within the station monitor ambient temperature and humidity in real time. Using these three parameters, the data processing unit can calculate the moisture evaporation intensity under current conditions—that is, the rate of moisture evaporation per unit area per unit time—based on established heat and mass transfer models (such as evaporation models based on saturated vapor pressure differentials).

[0057] Subsequently, the system invokes a preset attenuation compensation function. This function estimates the amount of water that may evaporate during the entire dielectric signal acquisition period based on the calculated water evaporation intensity and the time required for the measurement process, and dynamically compensates for this loss in the measurement results. One specific implementation involves introducing a compensation term into the mathematical model for calculating the moisture index using dielectric characteristic signals (such as resonant frequency offset). The magnitude of this compensation term is proportional to the water evaporation intensity and the measurement duration. Essentially, before reporting the final moisture index, the system states: "Based on environmental conditions, I estimate that 0.1% of the moisture was lost during the measurement; I'm adding it back now." This real-time, dynamic compensation ensures that the final moisture index more accurately reflects the original state of the pulverized coal in the process pipeline, avoiding systematic underestimation errors introduced by environmental changes.

[0058] Furthermore, the steps of extracting eigenvalues ​​from dielectric property signals, molecular vibrational spectral signals, and atomic characteristic spectral signals, and substituting them into a pre-set quantitative correlation equation set to simultaneously solve for moisture content, ash content, and calorific value include: The resonant frequency offset and quality factor are extracted from the dielectric property signal as dielectric characteristic values. The absorbance or reflectance at a preset wavelength position is extracted from the molecular vibrational spectral signal as a molecular vibrational characteristic value. The characteristic peak intensities of preset mineral elements are extracted from the atomic characteristic spectrum signal and used as the characteristic values ​​of the atomic characteristic spectrum. Based on the characteristic peak intensity of preset mineral elements, the coal type classification information of the current coal powder sample is identified; Retrieve a set of target quantitative correlation equations that match the coal type classification information from a pre-defined equation database; The extracted resonant frequency shift, quality factor, absorbance or reflectance, and characteristic peak intensity are substituted into the target quantitative correlation equations for solution to obtain moisture index, ash index, and calorific value index.

[0059] The core innovation of this solution lies in the fact that instead of using a one-size-fits-all model to analyze all coal samples, it first identifies the coal sample and then calls upon a customized analysis model tailored to it.

[0060] First, in the feature extraction stage, the system extracts the most effective information from the three types of signals acquired. For example, from the dielectric signal, it extracts not only the resonant frequency shift directly related to moisture content, but also the quality factor (Q value) reflecting the total ion concentration in the sample. Combining these two aspects allows for more accurate moisture analysis. From the molecular vibrational spectrum (near-infrared spectroscopy), the system focuses on absorbance at several specific wavelengths, such as 1450 nm and 1940 nm (related to the OH bonds in water) and multiple points in the 2200-2400 nm range (related to the CH, CO, and other chemical bonds in organic matter, and associated with calorific value). From the atomic characteristic spectrum (laser-induced breakdown spectrum), the system extracts the peak intensities of characteristic spectral lines representing the main constituent elements of ash (such as Si, Al, Fe, Ca, K, Mg).

[0061] The next crucial step is coal type identification. Coal from different origins and with varying degrees of coalification often possesses unique fingerprint characteristics in terms of the types and relative proportions of mineral elements it contains. For example, coal from certain regions may be rich in calcium, while coal from other regions may have higher iron and aluminum content. The system leverages this by extracting the characteristic peak intensities of multiple mineral elements as a vector and inputting it into a pre-trained classifier model (such as a support vector machine or decision tree). This classifier, having learned from a large amount of atomic feature spectrum data of known coal types during its development phase, can automatically identify the most likely coal type for the current sample based on the input elemental fingerprint and output coal type classification information, such as Datong bituminous coal from Shanxi or Dongsheng lignite from Inner Mongolia.

[0062] Once the coal type is identified, the system retrieves a set of quantitative correlation equations from a vast database of equations that precisely match the coal type's classification information. This database stores dozens or even hundreds of different sets of equations, each optimized for a specific coal type through analysis and modeling of a large number of samples, thus possessing extremely high specificity and accuracy.

[0063] Finally, the system substitutes all the feature values ​​extracted in this round of measurements (dielectric feature values, molecular vibrational feature values, and atomic spectrum feature values) into the newly selected set of quantitative correlation equations most suitable for the current coal sample, and solves them to obtain the final moisture, ash, and calorific value indices. This adaptive modeling strategy of identifying first and then analyzing greatly improves the method's adaptability to different coal sources and the accuracy of the analysis results.

[0064] Furthermore, the steps of comparing the obtained moisture, ash, and calorific value indices with the corresponding preset benchmark values ​​to generate and execute the coal mill adjustment command include: When the moisture index is higher than the preset moisture benchmark value, the adjustment command to reduce the coal feed rate of the coal mill and increase the drying air volume is executed. When the ash content index is higher than the preset ash content benchmark value, the adjustment command for blowing soot onto the boiler heating surface is executed. When the calorific value is lower than the preset calorific value benchmark, an adjustment command is executed to increase the output of the coal mill or increase the primary air temperature.

[0065] This series of instructions constitutes a feedforward control strategy based on real-time coal quality data, and its underlying logic is clear and consistent with combustion engineering principles. For example, when the system detects that the moisture content of the pulverized coal entering the furnace is higher than the preset moisture benchmark value (e.g., 10%), it means that the pulverized coal is too wet, making it difficult to grind and ignite. In response, the system will immediately send two coordinated instructions to the power plant's DCS system: first, reduce the coal feed rate of the pulverizer, i.e., slow down the coal feeding speed, allowing the coal to have a longer residence time in the pulverizer to be dried by hot air; second, increase the drying air volume or increase the air temperature to provide stronger drying capacity.

[0066] When the system detects that the ash content exceeds the preset ash content benchmark (e.g., 25%), it means that there are too many non-combustible minerals in the coal. These minerals form ash after combustion, which easily adheres to the boiler's water-cooled walls, superheaters, and other heating surfaces, forming coke or ash buildup. This severely hinders heat transfer and reduces boiler efficiency. Therefore, the system will issue an early warning and automatically execute the boiler heating surface soot blowing adjustment command, activating the steam soot blower or sonic soot blower to remove the ash before it forms a stubborn layer, maintaining the boiler's cleanliness and high efficiency.

[0067] When the system detects that the calorific value is lower than the preset benchmark value (e.g., 20 MJ / kg), it means that the energy density of the pulverized coal is insufficient. In order to maintain the boiler's evaporation rate and power generation load, more total heat must be supplied to the furnace per unit time. In response, the system will select to execute commands to increase the output of the coal mill (i.e., increase the speed or pressure of the coal mill to grind more pulverized coal) or increase the primary air temperature (i.e., use hotter air at a higher temperature to transport and heat the pulverized coal, so that it ignites more quickly and burns more completely, thereby improving combustion efficiency) based on the boiler's current load and operating strategy.

[0068] By combining real-time coal quality analysis results with this set of clear and automated control rules, the technical solution of this application ultimately achieves intelligent, precise, and forward-looking control of the combustion process.

[0069] Secondly, see Figure 2 This application also provides an intelligent analysis system for coal quality testing data, used to perform the intelligent analysis method for coal quality testing data described in any of the preceding claims. The system includes: The sampling module 210 is used to perform cyclic sampling operations at multiple spatial locations in the coal mill outlet pipe to obtain coal powder samples; The sample preparation module 220 is used to quantitatively extract coal powder samples and apply a preset pressure to form the test object; The signal acquisition module 230 is used to perform multi-physical quantity signal acquisition on the entity under test, and obtain dielectric property signals reflecting moisture content, molecular vibration spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. The index solving module 240 is used to extract the characteristic values ​​from the dielectric property signal, molecular vibrational spectrum signal and atomic characteristic spectrum signal, and substitute them into the preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index and calorific value index. The control execution module 250 is used to compare the obtained moisture index, ash index and calorific value index with the corresponding preset benchmark values, and generate and execute the coal mill adjustment command.

[0070] The surface defect compensation module is used to perform a topography scan on the surface of the entity under test before performing multi-physical quantity signal acquisition, so as to obtain surface integrity characteristics. When the surface integrity feature is lower than a preset threshold, a morphology correction coefficient is generated based on the distribution density of surface defects in the entity under test. After acquiring dielectric property signals reflecting moisture content, molecular vibrational spectral signals reflecting organic components, and atomic characteristic spectral signals reflecting mineral elements, the acquired molecular vibrational spectral signals are numerically compensated using morphology correction coefficients to offset diffuse reflection loss.

[0071] By dividing the functions into modules such as sampling, sample preparation, signal acquisition, index solving, and control execution, the hardware composition and functional architecture of the system are clarified, providing a clear implementation blueprint for the engineering and industrial application of this technology.

[0072] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An intelligent analysis method for coal quality testing data, characterized in that, The method includes: Cyclic sampling was performed at multiple spatial locations in the coal mill outlet pipe to obtain pulverized coal samples; The coal powder sample is quantitatively extracted and a preset pressure is applied to form a test object; Multi-physical quantity signal acquisition is performed on the entity to be tested to obtain dielectric property signals reflecting moisture content, molecular vibrational spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. The characteristic values ​​of the dielectric property signal, the molecular vibrational spectrum signal and the atomic characteristic spectrum signal are extracted and substituted into a preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index and calorific value index. The obtained moisture, ash, and calorific value indices are compared with the corresponding preset benchmark values ​​to generate and execute mill adjustment commands.

2. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, The step of performing cyclic sampling at multiple spatial locations at the coal mill outlet to obtain coal powder samples includes: Sampling channels distributed at multiple spatial locations in the coal mill outlet pipe are opened in turn according to preset time intervals. A negative pressure environment is established in the sampling channel to guide the coal powder airflow in the coal mill outlet pipe to the separation path, and the physical separation of coal powder particles in the coal powder airflow from the carrier gas is achieved by centrifugal force to obtain the coal powder sample.

3. The intelligent analysis method for coal quality testing data according to claim 2, characterized in that, The step of establishing a negative pressure environment in the sampling channel to guide the coal powder airflow in the coal mill outlet pipe to the separation path, and using centrifugal force to physically separate the coal powder particles in the coal powder airflow from the carrier gas to obtain the coal powder sample includes: The coal powder airflow is guided to the primary separation zone, and the first centrifugal force field is used to capture the first part of the coal powder particles and make them fall into the collection zone. The coal powder airflow after passing through the primary separation zone is guided to the secondary separation zone for secondary separation, where a second centrifugal force field is used to capture fine dust and cause it to fall into the collection area. The coal powder particles in the first part of the collection area are combined with the fine dust to obtain the coal powder sample.

4. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, The step of quantitatively extracting the coal powder sample and applying a preset pressure to form a test entity includes: A sample of coal powder of a predetermined weight is taken, a predetermined pressure is applied and maintained for a predetermined time to obtain the test entity with a predetermined saturation density.

5. The intelligent analysis method for coal quality testing data according to claim 4, characterized in that, The step of applying a preset pressure and maintaining it for a preset duration includes: Real-time monitoring of the correlation curve between pressure intensity and displacement change during the pressure application process; The elastic characteristics of the pulverized coal sample are determined based on the correlation curve. When the elastic feature meets the preset rebound condition, the preset duration is extended to eliminate internal stress.

6. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, Prior to the step of performing multi-physical quantity signal acquisition on the entity under test, the following is included: A topography scan is performed on the surface of the entity under test to obtain surface integrity features; When the surface integrity feature is lower than a preset threshold, a morphology correction coefficient is generated based on the distribution density of surface defects in the entity under test. After acquiring the dielectric property signal reflecting moisture content, the molecular vibrational spectral signal reflecting organic components, and the atomic characteristic spectral signal reflecting mineral elements, the process includes: The morphology correction coefficient is used to numerically compensate the acquired molecular vibrational spectral signal to offset diffuse reflection loss.

7. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, The step of performing multi-physical quantity signal acquisition on the entity under test within the same processing cycle includes: The initial surface temperature, ambient temperature, and ambient humidity of the object under test are obtained. The water evaporation intensity is calculated based on the difference between the initial surface temperature and the ambient temperature, as well as the ambient humidity. Based on the moisture evaporation intensity, a preset attenuation compensation function is invoked to increase the moisture response weight in the dielectric characteristic signal in real time, thereby obtaining the compensated dielectric characteristic signal.

8. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, The steps of extracting the characteristic values ​​from the dielectric property signal, the molecular vibrational spectrum signal, and the atomic characteristic spectrum signal, and substituting them into a preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index, and calorific value index include: The resonant frequency offset and quality factor are extracted from the dielectric property signal as dielectric characteristic values. The absorbance or reflectance at a preset wavelength position is extracted from the molecular vibrational spectral signal as a molecular vibrational characteristic value; The characteristic peak intensities of preset mineral elements are extracted from the atomic characteristic spectrum signal and used as the characteristic values ​​of the atomic characteristic spectrum. Based on the characteristic peak intensity of the preset mineral elements, the coal type classification information of the current coal powder sample is identified; Retrieve a set of target quantitative correlation equations that match the coal type classification information from a pre-defined equation database; The extracted resonant frequency offset and quality factor, absorbance or reflectance, and characteristic peak intensity are substituted into the target quantitative correlation equations for solution to obtain the moisture index, ash index, and calorific value index.

9. The intelligent analysis method for coal quality testing data according to claim 1, characterized in that, The step of comparing the obtained moisture, ash, and calorific value indices with corresponding preset benchmark values ​​to generate and execute the coal mill adjustment command includes: When the moisture index is higher than the preset moisture benchmark value, an adjustment command is executed to reduce the coal feed rate of the coal mill and increase the drying air volume. When the ash content index is higher than the preset ash content benchmark value, the adjustment command for blowing soot onto the boiler heating surface is executed; When the calorific value index is lower than the preset calorific value benchmark, an adjustment command is executed to increase the output of the coal mill or increase the primary air temperature.

10. A coal quality testing data intelligent analysis system, used to execute the coal quality testing data intelligent analysis method as described in any one of claims 1 to 9, characterized in that, The system includes: The sampling module is used to perform cyclic sampling operations at multiple spatial locations in the coal mill outlet pipe to obtain coal powder samples; The sample preparation module is used to quantitatively extract the coal powder sample and apply a preset pressure to form the test object; The signal acquisition module is used to perform multi-physical quantity signal acquisition on the entity under test to obtain dielectric property signals reflecting moisture content, molecular vibration spectrum signals reflecting organic components, and atomic characteristic spectrum signals reflecting mineral elements. The index solving module is used to extract the feature values ​​from the dielectric property signal, the molecular vibrational spectrum signal and the atomic feature spectrum signal, and substitute them into the preset quantitative correlation equation set to simultaneously solve for the moisture index, ash index and calorific value index. The control execution module is used to compare the obtained moisture index, ash index and calorific value index with the corresponding preset benchmark values, and generate and execute the coal mill adjustment command.

Images