Pulverizing system fine control method based on pulverized coal flow characteristics

Abstract

The invention relates to a coal pulverizing system fine control method based on pulverized coal flow characteristics. The method comprises the following steps: firstly, acquiring an original sound wave signal in the flowing process of pulverized coal, and performing spectral decomposition and feature extraction adjustment on the original sound wave signal to obtain a real-time change trend of pulverized coal fineness distribution; then obtaining the electrified signal intensity of the pulverized coal particles and combining with the real-time change trend for analysis to obtain an accurate measurement value of the pulverized coal flow velocity; then, inputting the real-time change trend and the accurate measurement value of the pulverized coal flow velocity into a particle kinematics model, and calculating to obtain a dynamic estimation value of the pulverized coal concentration; the real-time change trend, the accurate measurement value of the pulverized coal flow velocity and the dynamic estimation value of the pulverized coal concentration are fused, and comprehensive monitoring parameters of pulverized coal flow are generated; and finally, dynamically adjusting a control instruction in the powder making process according to the comprehensive monitoring parameters, and finally obtaining multi-parameter monitoring data. The powder making efficiency and powder making refinement are effectively improved, and the stability and high efficiency of operation of a powder making system are guaranteed.

Description

Technical Field

[0001] The present invention belongs to the field of electric power energy, and particularly relates to a fine control method for a coal pulverizing system based on the flow characteristics of pulverized coal. Background Art

[0002] With the development of technologies in the field of electric power energy, a fine control technology for a coal pulverizing system based on the flow characteristics of pulverized coal has emerged. In the field of thermal power generation, the operating conditions of the coal pulverizing system play a decisive role in power generation efficiency, cost control, and environmental protection indicators. As the pulverized coal used as the boiler fuel, its quality and flow characteristics directly affect the combustion effect. If the fineness, flow rate, and concentration of the pulverized coal are not properly controlled, it will not only lead to incomplete combustion and reduce the boiler thermal efficiency, but also increase pollutant emissions and cause environmental pollution problems. At the same time, the complex structure and harsh working environment of the coal pulverizing system result in problems such as slow load response and poor deep regulation ability during operation, and phenomena such as bin blocking and coal collapse accidents also occur from time to time. Under this background, conventional control methods are difficult to meet the requirements of modern production for the stability, efficiency, and refinement of the coal pulverizing system. Summary of the Invention

[0003] Based on this, it is necessary to provide a fine control method for a coal pulverizing system based on the flow characteristics of pulverized coal that can meet the requirements of modern production for the stability, efficiency, and refinement of the coal pulverizing system in view of the above technical problems.

[0004] In a first aspect, the present application provides a fine control method for a coal pulverizing system based on the flow characteristics of pulverized coal, including:

[0005] Obtain multi-band original acoustic signals in the flow of pulverized coal; perform spectral decomposition and feature extraction adjustment on the original acoustic signals to obtain the real-time change trend of the fineness distribution of pulverized coal.

[0006] Obtain the charged signal intensity of pulverized coal particles, analyze the multi-point electrostatic field intensity in combination with the real-time change trend, and obtain an accurate measurement value of the flow rate of pulverized coal.

[0007] Input the real-time change trend and the accurate measurement value into a trained particle kinematics model to deduce the mass distribution characteristics of pulverized coal in the pipeline, and calculate the dynamic estimated value of the pulverized coal concentration.

[0008] Fuse the real-time change trend, the accurate measurement value, and the dynamic estimated value of the concentration to generate a comprehensive monitoring parameter for the flow of pulverized coal.

[0009] Dynamically adjust the control instructions for the coal pulverizing process according to the comprehensive monitoring parameter and perform adaptive calibration on the operating state of the sensor to obtain stable multi-parameter monitoring data for the flow of pulverized coal.

[0010] In one embodiment, the original acoustic wave signal is subjected to spectral decomposition and feature extraction adjustment to obtain the real-time change trend of the pulverized coal fineness distribution, including:

[0011] The original acoustic wave signal is subjected to frequency-domain conversion to obtain the energy distribution data of each frequency band; the original acoustic wave signal includes the acoustic wave characteristics of particle collisions with different particle sizes.

[0012] According to the mapping relationship between the pulverized coal particle size and the acoustic wave frequency band, the dynamic clustering algorithm is used to decouple the features of the energy distribution data of each frequency band, and the frequency band energy weight coefficient related to the pulverized coal fineness is separated.

[0013] Based on the frequency band energy weight coefficient, a frequency band energy change matrix is constructed; the dimension of the frequency band energy change matrix corresponds to the pulverized coal particle size distribution interval.

[0014] The frequency band energy change matrix is input into the trained pulverized coal fineness regression model to obtain the particle size cumulative distribution function of the pulverized coal particles.

[0015] According to the particle size cumulative distribution function, the real-time distribution curve of the pulverized coal fineness is calculated to obtain the real-time change trend of the pulverized coal fineness distribution.

[0016] In one embodiment, according to the mapping relationship between the pulverized coal particle size and the acoustic wave frequency band, the dynamic clustering algorithm is used to decouple the features of the energy distribution data of each frequency band, and the frequency band energy weight coefficient related to the pulverized coal fineness is separated, including:

[0017] Obtain the mapping relationship data between the pulverized coal particle size and the acoustic wave frequency band; the mapping relationship data includes the frequency band division rules corresponding to different particle sizes.

[0018] According to the frequency band division rules, the mapping relationship data is decomposed into frequency bands to generate an energy distribution matrix for each frequency band.

[0019] The dynamic clustering algorithm is used to perform iterative calculations on the energy distribution matrix to obtain the sample attribution category vector.

[0020] According to the sample attribution category vector, the characteristic frequency band index value related to the pulverized coal fineness is extracted.

[0021] Based on the characteristic frequency band index value, the target frequency band energy sequence is intercepted, and the proportion parameter of the target frequency band energy sequence in the total energy is calculated.

[0022] According to the proportion parameter, a frequency band weight optimization model is established using a linear mapping relationship to obtain the frequency band energy weight coefficient.

[0023] In one embodiment, the charged signal intensity of the pulverized coal particles is obtained, and the multi-point electrostatic field intensity is analyzed in combination with the real-time change trend to obtain an accurate measurement value of the pulverized coal flow rate, including:

[0024] Generate a time - series feature matrix based on the strength of the charged signal; the time - series feature matrix includes the electric - field fluctuation amplitude and phase difference of each monitoring point.

[0025] Input the time - series feature matrix into a dynamic calibration model to extract the signal - noise components and obtain the high - frequency interference pattern; the dynamic calibration model separates the signal - noise components based on the sliding - window mechanism.

[0026] Perform spatial matching between the high - frequency interference pattern and the electrostatic - field distribution parameters to obtain the flow - velocity compensation coefficient; the flow - velocity compensation coefficient is used to correct the initial flow - velocity value.

[0027] Use a regression algorithm to fuse the flow - velocity compensation coefficient and the initial flow - velocity value to obtain the accurate measurement value of the pulverized - coal flow velocity.

[0028] In one embodiment, performing spatial matching between the high - frequency interference pattern and the electrostatic - field distribution parameters to obtain the flow - velocity compensation coefficient includes:

[0029] Generate grid alignment parameters for the electrostatic - field distribution parameters according to the frequency - feature set of the high - frequency interference pattern; the frequency - feature set includes the interference - intensity distribution of multiple spatially discrete points.

[0030] Interpolate and reconstruct the electrostatic - field distribution parameters based on the grid alignment parameters to obtain the electrostatic - field gradient matrix after spatial matching.

[0031] Input the electrostatic - field gradient matrix and the frequency - feature set into a compensation model to obtain the flow - velocity compensation coefficient with weight factors.

[0032] In one embodiment, inputting the real - time change trend and the accurate measurement value into a trained particle kinematics model to deduce the mass - distribution characteristics of pulverized coal in the pipeline and calculate the dynamic estimated value of the pulverized - coal concentration, including:

[0033] Obtain the flow - velocity fluctuation data and pressure - gradient data in the real - time change trend.

[0034] Generate a flow - velocity distribution matrix of the pipeline cross - section according to the flow - velocity fluctuation data.

[0035] According to the coupling relationship between the pressure - gradient data and the flow - velocity distribution matrix, use a multi - physical - field coupling algorithm to calculate the set of migration trajectories of pulverized - coal particles.

[0036] Input the set of migration trajectories into a trained particle kinematics model to obtain the probability - density function of the migration trajectories.

[0037] Perform three - dimensional grid - based numerical integration according to the probability - density function to obtain the spatial discreteness distribution of the pulverized - coal mass.

[0038] Extract the mass fraction coefficient of each grid cell for the spatial dispersion distribution, and generate a dynamic concentration gradient map by combining the vector components of the flow velocity distribution matrix.

[0039] Among them, if there is a mutation in the mass fraction coefficient in a local area of the dynamic concentration gradient map, update the weight parameter of the probability density function to obtain the updated weight parameter.

[0040] Recalculate the set of migration trajectories according to the updated weight parameter to generate a corrected dynamic concentration gradient map.

[0041] Compare the residual of the corrected dynamic concentration gradient map with the pressure gradient data. If the residual value exceeds the preset threshold, adjust the boundary condition parameters of the multi-physics coupling algorithm according to the residual comparison result to obtain the dynamic estimated value of the final pulverized coal concentration.

[0042] In one embodiment, the probability density function is calculated by the following formula:

[0043]

[0044]

[0045] Among them, f(x) represents the probability density function of the migration trajectory, h represents the bandwidth, n represents the number of trajectories included in the set of migration trajectories, K(x, y) represents the kernel function, x, y represent vectors in d-dimensional space, d represents the spatial dimension (in the pulverized coal pipeline scenario, generally d = 3), σ represents the bandwidth parameter, X i (t) represents the i-th migration trajectory of the pulverized coal particles in the set of migration trajectories at time t.

[0046] In one embodiment, dynamically adjust the control instructions for the coal pulverizing process according to the comprehensive monitoring parameters and adaptively calibrate the operating state of the sensor to obtain stable multi-parameter monitoring data of the pulverized coal flow, including:

[0047] Obtain the fluctuation value of the pulverized coal flow velocity in the comprehensive monitoring parameters.

[0048] Calculate according to the fluctuation value of the pulverized coal flow velocity using the correlation model between the pulverized coal concentration and the pressure gradient to obtain the dynamic adjustment coefficient.

[0049] Input the dynamic adjustment coefficient into the coal pulverizing control instruction to obtain the real-time control parameters including the valve opening and the mill speed.

[0050] Calculate the instruction difference value between the real-time control parameter and the original data of the sensor. If the instruction difference value exceeds the preset threshold, call the adaptive compensation model to correct the sensor drift amount to obtain the corrected sensor parameter; the adaptive compensation model is trained and generated based on the historical data of the temperature gradient and the pulverized coal concentration.

[0051] Fuse the corrected sensor parameters with the real-time control parameters to obtain stable multi-parameter monitoring data of pulverized coal flow.

[0052] In a second aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the foregoing method is implemented.

[0053] In a third aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the foregoing method is implemented.

[0054] The above-mentioned refined control method for a coal pulverizing system based on the characteristics of pulverized coal flow first obtains multi-band original acoustic signals in the pulverized coal flow, performs spectral decomposition and feature extraction adjustment on them to obtain the real-time change trend of the pulverized coal fineness distribution; then obtains the charged signal intensity of the pulverized coal particles, analyzes the multi-point electrostatic field intensity in combination with the above real-time change trend to obtain an accurate measurement value of the pulverized coal flow rate; subsequently, inputs the real-time change trend of the pulverized coal fineness distribution and the accurate measurement value of the pulverized coal flow rate into a trained particle kinematics model to deduce the mass distribution characteristics of the pulverized coal in the pipeline and calculate the dynamic estimated value of the pulverized coal concentration; then fuses the real-time change trend of the pulverized coal fineness distribution, the accurate measurement value of the pulverized coal flow rate, and the dynamic estimated value of the pulverized coal concentration to generate comprehensive monitoring parameters of the pulverized coal flow; finally, dynamically adjusts the control instructions for the coal pulverizing process according to the comprehensive monitoring parameters, and adaptively calibrates the operating state of the sensor to finally obtain stable multi-parameter monitoring data of the pulverized coal flow. It effectively improves the pulverizing efficiency and the refinement of pulverized coal quality, ensures the stability and reliability of the operation of the coal pulverizing system, reduces the risk of equipment failure and maintenance costs, and also helps to improve the boiler combustion efficiency, reduce energy waste and pollutant emissions, and achieve high efficiency, energy conservation and environmental protection in the coal pulverizing link of the thermal power generation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0056] Figure 1 It is a flowchart of the refined control method for a coal pulverizing system based on the characteristics of pulverized coal flow provided by an embodiment of the present invention;

[0057] Figure 2A flowchart for dynamically adjusting control instructions for the coal pulverization process based on comprehensive monitoring parameters and adaptively calibrating the operating state of sensors to obtain stable multi-parameter monitoring data of coal powder flow provided by an embodiment of the present invention. Detailed implementation manners

[0058] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0059] In one of the embodiments, as Figure 1 shown, the present application provides a refined control method for a coal pulverization system based on the coal powder flow characteristics, which may include the following steps:

[0060] Step S101, obtaining multi-band original acoustic wave signals in the coal powder flow; performing spectral decomposition and feature extraction adjustment on the original acoustic wave signals to obtain the real-time change trend of the coal powder fineness distribution.

[0061] During the operation of the coal pulverization system, the coal powder flow will generate multi-band original acoustic wave signals. These signals can be effectively obtained through acoustic wave sensors installed at specific positions in the coal powder pipeline. These sensors have high sensitivity and can accurately capture weak acoustic waves. After obtaining the signals, professional spectral analysis technology is used to perform spectral decomposition on the original acoustic wave signals. This process can disassemble complex acoustic wave signals according to frequency components, clearly presenting the energy distribution of different frequency bands. Then, according to the internal mapping relationship between the coal powder particle size and the acoustic wave frequency band, a feature extraction algorithm is used to perform feature extraction adjustment on the decomposed signals. By analyzing the change rules of the energy of different frequency bands, key feature information related to the coal powder fineness is determined, and finally, the real-time trend that can accurately reflect the change of the coal powder fineness distribution over time is obtained.

[0062] Step S102, obtaining the charged signal intensity of the coal powder particles and analyzing the multi-point electrostatic field intensity in combination with the real-time change trend to obtain an accurate measurement value of the coal powder flow rate.

[0063] Specifically, the principle of electrostatic induction is utilized to obtain the charged signal intensity of pulverized coal particles through an electrostatic sensor. Since pulverized coal will become electrostatically charged due to friction and other reasons during the flow process, these electrostatic signals contain information related to the flow state of pulverized coal. The obtained charged signal intensity data is combined with the real-time change trend of the pulverized coal fineness distribution, because the change in pulverized coal fineness may affect the charging characteristics of pulverized coal and the electrostatic field distribution. On this basis, the electrostatic field intensity at multiple monitoring points is analyzed in depth, and signal processing algorithms are used to remove interference signals and extract effective information. By calculating parameters such as the change in electric field intensity and the signal propagation time difference, and using a specific flow velocity calculation model, after multiple iterations and calibrations, an accurate measurement value of the pulverized coal flow velocity with high precision is obtained.

[0064] Step S103: Input the real-time change trend and the accurate measurement value into the trained particle kinematics model to deduce the mass distribution characteristics of pulverized coal in the pipeline, and calculate the dynamic estimated value of the pulverized coal concentration.

[0065] Take the real-time change trend of the pulverized coal fineness distribution and the accurate measurement value of the obtained pulverized coal flow velocity as input data, and import them into the pre-trained particle kinematics model. This model is constructed based on the physical laws of the movement of pulverized coal particles in the pipeline, fully considering factors such as the force on pulverized coal particles, collision behavior, and interaction with the pipeline wall. Through the analysis of these input data and the operation of the model, the movement trajectory and diffusion situation of pulverized coal in the pipeline are simulated, and then the mass distribution characteristics of pulverized coal in the pipeline are deduced. According to the model output results, corresponding mathematical calculation methods, such as integral operation, statistical analysis, etc., are used to calculate the dynamic estimated value that can reflect the change in the concentration of pulverized coal at different positions in the pipeline in real time.

[0066] Step S104: Integrate the real-time change trend, the accurate measurement value, and the dynamic estimated value of the concentration to generate comprehensive monitoring parameters for the pulverized coal flow.

[0067] Fuse the three groups of key data: the real-time change trend of the pulverized coal fineness distribution, the accurate measurement value of the pulverized coal flow velocity, and the calculated dynamic estimated value of the pulverized coal concentration. Adopt advanced data fusion algorithms, such as weighted fusion, Kalman filter fusion, etc., assign corresponding weights to each parameter according to its importance and reliability, organically combine different types of data, eliminate redundant and contradictory information between the data, and generate comprehensive monitoring parameters that can comprehensively and integrally reflect the pulverized coal flow state. These comprehensive monitoring parameters cover information in multiple aspects such as pulverized coal fineness, flow velocity, and concentration, providing a unified and intuitive quantitative index for the overall operation state assessment of the coal pulverizing system.

[0068] Step S105: Dynamically adjust the control instructions for the coal pulverization process based on the comprehensive monitoring parameters and adaptively calibrate the operating states of the sensors to obtain stable multi-parameter monitoring data of the pulverized coal flow.

[0069] Specifically, based on the generated comprehensive monitoring parameters, dynamically adjust the control instructions for the coal pulverization process. By analyzing the differences between the comprehensive monitoring parameters and the ideal operating state parameters of the coal pulverization system, use control algorithms to calculate the adjustment amounts of the control instructions, such as adjusting the rotation speed of the coal mill, the coal feeding amount of the coal feeder, the air volume of the fan, etc., so that the coal pulverization system can adjust the operating parameters in a timely manner according to the actual pulverized coal flow state and maintain the best operating conditions. At the same time, adaptively calibrate the operating states of the acoustic wave sensors and electrostatic sensors. According to the changes in the comprehensive monitoring parameters and the historical data of the sensors, judge whether there are problems such as drift and faults in the sensors. If abnormalities are found, automatically adjust parameters such as the sensitivity and calibration coefficient of the sensors, or conduct fault diagnosis and repair on the sensors to ensure that the sensors can continuously and stably obtain accurate monitoring data, and finally obtain stable and reliable multi-parameter monitoring data of the pulverized coal flow, ensuring the long-term, stable and efficient operation of the coal pulverization system.

[0070] The above-mentioned refined control method for the coal pulverization system based on the pulverized coal flow characteristics first obtains the multi-band original acoustic wave signals in the pulverized coal flow, performs spectral decomposition and feature extraction adjustment on them to obtain the real-time change trend of the pulverized coal fineness distribution; then obtains the charged signal intensity of the pulverized coal particles, combines the above real-time change trend to analyze the multi-point electrostatic field intensity, and obtains the accurate measurement value of the pulverized coal flow rate; subsequently, inputs the real-time change trend of the pulverized coal fineness distribution and the accurate measurement value of the pulverized coal flow rate into the trained particle kinematics model to deduce the mass distribution characteristics of the pulverized coal in the pipeline and calculate the dynamic estimated value of the pulverized coal concentration; then fuse the real-time change trend of the pulverized coal fineness distribution, the accurate measurement value of the pulverized coal flow rate, and the dynamic estimated value of the pulverized coal concentration to generate the comprehensive monitoring parameters of the pulverized coal flow; finally, dynamically adjust the control instructions for the coal pulverization process based on the comprehensive monitoring parameters, and at the same time adaptively calibrate the operating states of the sensors to finally obtain stable multi-parameter monitoring data of the pulverized coal flow. It effectively improves the pulverization efficiency and the refinement of the pulverized coal quality, ensures the stability and reliability of the operation of the coal pulverization system, reduces the equipment failure risk and maintenance cost, and also helps to improve the boiler combustion efficiency, reduce energy waste and pollutant emissions, and realize the high efficiency, energy conservation and environmental protection of the coal pulverization link in the thermal power generation process.

[0071] In one embodiment, performing spectral decomposition and feature extraction adjustment on the original acoustic wave signals to obtain the real-time change trend of the pulverized coal fineness distribution may include the following steps:

[0072] Step S201: Perform frequency-domain conversion on the original acoustic wave signal to obtain the energy distribution data of each frequency band. The original acoustic wave signal includes the acoustic wave characteristics of particle collisions with different particle sizes.

[0073] Step S202: Use the dynamic clustering algorithm to perform feature decoupling on the energy distribution data of each frequency band according to the mapping relationship between the pulverized coal particle size and the acoustic wave frequency band, and separately obtain the frequency band energy weight coefficients related to the pulverized coal fineness.

[0074] Step S203: Construct a frequency band energy change matrix based on the frequency band energy weight coefficients. The dimension of the frequency band energy change matrix corresponds to the pulverized coal particle size distribution interval.

[0075] Step S204: Input the frequency band energy change matrix into the trained pulverized coal fineness regression model to obtain the particle size cumulative distribution function of the pulverized coal particles.

[0076] Step S205: Calculate the real-time distribution curve of the pulverized coal fineness according to the particle size cumulative distribution function to obtain the real-time change trend of the pulverized coal fineness distribution.

[0077] After obtaining the original acoustic wave signal generated by the pulverized coal flow, first perform frequency-domain conversion on it. This process uses mathematical methods such as Fourier transform to convert the original acoustic wave signal from the time domain to the frequency domain, and then obtain the energy distribution data of each frequency band. The original acoustic wave signal contains the acoustic wave characteristics generated by particle collisions with different particle sizes. When pulverized coal particles with different particle sizes collide with other objects, they will generate acoustic waves with specific frequency characteristics. Based on the mapping relationship between the pulverized coal particle size and the acoustic wave frequency band, use the dynamic clustering algorithm to perform feature decoupling on the energy distribution data of each frequency band. The dynamic clustering algorithm can automatically identify and divide different categories according to the internal characteristics of the data. Through this algorithm, the frequency band energy related to the pulverized coal fineness is separated to obtain the frequency band energy weight coefficients. Then, construct a frequency band energy change matrix based on these weight coefficients. The dimension of this matrix corresponds to the pulverized coal particle size distribution interval, and can intuitively reflect the energy change of the pulverized coal in different particle size intervals. Subsequently, input the frequency band energy change matrix into the pre-trained pulverized coal fineness regression model, and the model will perform calculations according to the input data and output the particle size cumulative distribution function of the pulverized coal particles. Finally, based on the particle size cumulative distribution function, draw the real-time distribution curve of the pulverized coal fineness through specific calculation methods, so as to obtain the real-time change trend of the pulverized coal fineness distribution.

[0078] On the one hand, it enables operators to understand the dynamic changes in the fineness of pulverized coal in real time, promptly detect abnormal fluctuations in the fineness of pulverized coal, and then take corresponding measures for adjustment to ensure the stability of the quality of pulverized coal. This helps to improve the combustion efficiency of the boiler, reduce the phenomenon of incomplete combustion caused by improper fineness of pulverized coal, and reduce energy consumption and pollutant emissions. On the other hand, it provides a reliable basis for the automatic control of the coal pulverizing system, facilitates the realization of refined and intelligent control of the coal pulverizing process, improves the overall operating efficiency and reliability of the coal pulverizing system, and enhances the stability and economic benefits of the thermal power generation production process.

[0079] In one embodiment, according to the mapping relationship between the particle size of pulverized coal and the acoustic frequency band, the dynamic clustering algorithm is used to perform feature decoupling on the energy distribution data of each frequency band, and the frequency band energy weight coefficient related to the fineness of pulverized coal can be separated, which may include the following steps:

[0080] Step S301, obtain the mapping relationship data between the particle size of pulverized coal and the acoustic frequency band; the mapping relationship data includes the frequency band division rules corresponding to different particle sizes.

[0081] Step S302, perform frequency band decomposition on the mapping relationship data according to the frequency band division rules to generate the energy distribution matrix of each frequency band.

[0082] Step S303, use the dynamic clustering algorithm to perform iterative calculations on the energy distribution matrix to obtain the sample belonging category vector.

[0083] Step S304, extract the characteristic frequency band index value related to the fineness of pulverized coal according to the sample belonging category vector.

[0084] Step S305, intercept the target frequency band energy sequence based on the characteristic frequency band index value, and calculate the proportion parameter of the target frequency band energy sequence in the total energy.

[0085] Step S306, establish a frequency band weight optimization model according to the proportion parameter using a linear mapping relationship to obtain the frequency band energy weight coefficient.

[0086] Specifically, the mapping relationship data contains the frequency band division rules corresponding to different particle sizes of pulverized coal. Based on this rule, perform frequency band decomposition operations on the mapping relationship data to generate the energy distribution matrix of each frequency band. Then, use the dynamic clustering algorithm to perform iterative calculations on the energy distribution matrix, and then obtain the sample belonging category vector. Subsequently, based on the sample belonging category vector, extract the characteristic frequency band index value closely related to the fineness of pulverized coal. Based on this characteristic frequency band index value, intercept the target frequency band energy sequence and calculate the proportion parameter of the target frequency band energy sequence in the total energy. Finally, according to this proportion parameter, construct a frequency band weight optimization model by means of a linear mapping relationship, and thus obtain the frequency band energy weight coefficient.

[0087] This embodiment can effectively improve the accuracy and reliability of pulverized coal fineness monitoring. This method provides strong technical support for the quality control of pulverized coal in the industrial production process, helps optimize the production process, improve production efficiency, reduce production costs, and ensure the stability and safety of the production process.

[0088] In one embodiment, obtaining the charged signal intensity of pulverized coal particles and analyzing the multi-point electrostatic field intensity in combination with the real-time change trend to obtain an accurate measurement value of the pulverized coal flow rate may include the following steps:

[0089] Step S401, generating a time series feature matrix according to the charged signal intensity; the time series feature matrix includes the electric field fluctuation amplitude and phase difference of each monitoring point.

[0090] Step S402, inputting the time series feature matrix into a dynamic calibration model to extract the signal noise component and obtain a high-frequency interference pattern; the dynamic calibration model separates the signal noise component based on the sliding window mechanism.

[0091] Step S403, performing spatial matching on the high-frequency interference pattern and the electrostatic field distribution parameters to obtain a flow rate compensation coefficient; the flow rate compensation coefficient is used to correct the initial flow rate value.

[0092] Step S404, using a regression algorithm to fuse the flow rate compensation coefficient and the initial flow rate value to obtain an accurate measurement value of the pulverized coal flow rate.

[0093] Specifically, first, based on the obtained charged signal intensity of pulverized coal particles, by analyzing and processing the signal in chronological order, a time series feature matrix is generated. This matrix details key information such as the electric field fluctuation amplitude and phase difference of each monitoring point at different times, and these information contain the changes in the charged characteristics during the pulverized coal flow process. Then, the generated time series feature matrix is input into a dynamic calibration model constructed based on the sliding window mechanism. This model will analyze the data in the matrix segment by segment with the sliding window as the unit, and through setting reasonable thresholds and algorithm rules, accurately extract the noise components in the signal, and then obtain the high-frequency interference pattern. Subsequently, the high-frequency interference pattern is spatially matched with the pre-determined electrostatic field distribution parameters, and considering factors such as electric field strength, direction, and spatial position, the flow rate compensation coefficient is calculated. This coefficient reflects the influence degree of noise and interference on the initial flow rate measurement value and is used to correct the initial flow rate value. Finally, a regression algorithm is used to fuse the flow rate compensation coefficient and the initial flow rate value, comprehensively considering the weights and correlations of the two, and finally an accurate measurement value of the pulverized coal flow rate is obtained.

[0094] This embodiment effectively removes the noise interference in the signal and greatly improves the accuracy of the pulverized coal flow rate measurement. The accurate flow rate data provides a reliable basis for the operation monitoring and optimization of the coal pulverizing system, enabling the operator to timely and accurately grasp the flow state of the pulverized coal in the pipeline. The accurate flow rate measurement helps to achieve the refined control of the coal pulverizing process. For example, according to the accurate flow rate data, the operation parameters of equipment such as coal mills and coal feeders can be reasonably adjusted, the air-powder ratio can be optimized, the combustion efficiency can be improved, energy waste and equipment wear can be reduced, and thus the operation stability and economy of the entire coal pulverizing system can be enhanced.

[0095] In one embodiment, spatially matching the high-frequency interference pattern with the electrostatic field distribution parameters to obtain the flow rate compensation coefficient may include the following steps:

[0096] Step S501, generating grid alignment parameters of the electrostatic field distribution parameters according to the frequency feature set of the high-frequency interference pattern; the frequency feature set includes the interference intensity distribution of multiple spatially discrete points.

[0097] Step S502, performing interpolation reconstruction on the electrostatic field distribution parameters based on the grid alignment parameters to obtain the electrostatic field gradient matrix after spatial matching.

[0098] Step S503, inputting the electrostatic field gradient matrix and the frequency feature set into the compensation model to obtain the flow rate compensation coefficient with weight factors.

[0099] When dealing with the high-frequency interference problem, the grid alignment parameters of the electrostatic field distribution parameters can be generated according to the frequency feature set of the high-frequency interference pattern, where the frequency feature set covers the interference intensity distribution of multiple spatially discrete points. Based on the generated grid alignment parameters, interpolation reconstruction operations are carried out on the electrostatic field distribution parameters, and then the electrostatic field gradient matrix after spatial matching is obtained. After that, the electrostatic field gradient matrix and the frequency feature set are input into the compensation model together, and finally the flow rate compensation coefficient with weight factors is obtained.

[0100] By using the frequency feature set to generate grid alignment parameters and perform interpolation reconstruction, the matching degree between the electrostatic field distribution parameters and the actual space can be effectively improved, so that the obtained electrostatic field gradient matrix can more accurately reflect the actual situation. The flow rate compensation coefficient obtained by inputting it and the frequency feature set into the compensation model can accurately compensate the flow rate under high-frequency interference. For industrial production processes with high-frequency interference, such as the fields of electronic manufacturing and power transmission, it can significantly improve the system stability and measurement accuracy, reduce the influence of interference on production and measurement results, and ensure the efficient operation of the production process and the stable and reliable quality of products.

[0101] In one embodiment, the real-time change trend and the accurate measurement value are input into the trained particle kinematics model to deduce the mass distribution characteristics of pulverized coal in the pipeline, and the dynamic estimated value of the pulverized coal concentration can be obtained, which may include the following steps:

[0102] Step S601: Obtain the flow velocity fluctuation data and the pressure gradient data in the real-time change trend.

[0103] Step S602: Generate a flow velocity distribution matrix of the pipeline cross-section according to the flow velocity fluctuation data.

[0104] Step S603: According to the coupling relationship between the pressure gradient data and the flow velocity distribution matrix, use the multi-physical field coupling algorithm to calculate the set of migration trajectories of pulverized coal particles.

[0105] Step S604: Input the set of migration trajectories into the trained particle kinematics model to obtain the probability density function of the migration trajectories.

[0106] Step S605: Perform three-dimensional grid numerical integration according to the probability density function to obtain the spatial dispersion distribution of the pulverized coal mass.

[0107] Step S606: Extract the mass proportion coefficient of each grid unit from the spatial dispersion distribution, and combine the vector components of the flow velocity distribution matrix to generate a dynamic concentration gradient map.

[0108] Among them, if there is a sudden change in the mass proportion coefficient in a local area of the dynamic concentration gradient map, the weight parameter of the probability density function is updated to obtain the updated weight parameter.

[0109] Step S607: Recalculate the set of migration trajectories according to the updated weight parameter to generate a corrected dynamic concentration gradient map.

[0110] Step S608: Compare the residual of the corrected dynamic concentration gradient map with the pressure gradient data. If the residual value exceeds the preset threshold, adjust the boundary condition parameters of the multi-physical field coupling algorithm according to the residual comparison result to obtain the dynamic estimated value of the final pulverized coal concentration.

[0111] Specifically, first, the flow velocity fluctuation data and the pressure gradient data are extracted from the real-time change trend of the pulverized coal fineness distribution. The flow velocity fluctuation data reflects the change of the flow velocity of the pulverized coal in the pipeline, and the pressure gradient data reflects the change trend of the pressure in the pipeline. Then, using the flow velocity fluctuation data, a flow velocity distribution matrix of the pipeline cross-section is constructed through a specific algorithm, which intuitively presents the flow velocity information at different positions on the pipeline cross-section. Based on the coupling relationship between the pressure gradient data and the flow velocity distribution matrix, with the help of the multi-physics field coupling algorithm, comprehensively considering the interaction of various physical factors such as the pressure and flow velocity received by the pulverized coal particles, the set of migration trajectories of the pulverized coal particles in the pipeline is calculated. The set of these migration trajectories is input into a pre-trained particle kinematics model, and the model will analyze according to the movement law of the pulverized coal particles, and then obtain the probability density function of the migration trajectories. According to this probability density function, using the three-dimensional grid numerical integration method, the distribution of the pulverized coal in space is discretized to obtain the spatial dispersion distribution of the pulverized coal mass. Then, the mass proportion coefficient of each grid unit is extracted from the spatial dispersion distribution, and combined with the vector components of the flow velocity distribution matrix, a dynamic concentration gradient map is generated to visually display the distribution and change of the pulverized coal concentration in the pipeline. In this process, if there is a sudden change in the mass proportion coefficient in a local area of the dynamic concentration gradient map, the weight parameters of the probability density function are updated, and the set of migration trajectories is recalculated to generate a corrected dynamic concentration gradient map. Finally, the corrected dynamic concentration gradient map is compared with the pressure gradient data for residuals. If the residual value exceeds the preset threshold, the boundary condition parameters of the multi-physics field coupling algorithm are adjusted according to the residual comparison result. After repeated iterative calculations, the accurate dynamic estimation value of the pulverized coal concentration is finally obtained.

[0112] This embodiment can achieve accurate monitoring and dynamic estimation of the pulverized coal concentration. On the one hand, the accurate pulverized coal concentration data provides accurate pulverized coal distribution information for the operator, which helps to timely detect abnormal areas of the pulverized coal concentration and avoid problems such as incomplete combustion and pipeline blockage caused by uneven concentration, ensuring the safe and stable operation of the coal pulverizing system. On the other hand, based on these data, the operating parameters of the coal pulverizing system can be optimized and adjusted, such as reasonably adjusting the coal feeding amount of the coal feeder and the output of the coal mill, improving the utilization efficiency of the pulverized coal, reducing the production cost, enhancing the energy utilization efficiency and economic benefits of the entire thermal power generation process, and at the same time reducing environmental pollution and promoting the sustainable development of the energy industry.

[0113] In one of the embodiments, the probability density function can be calculated by the following formula:

[0114]

[0115] Among them, f(x) represents the probability density function of the migration trajectory, h represents the bandwidth, n represents the number of trajectories included in the migration trajectory set, K(x, y) represents the kernel function, x and y represent vectors in the d-dimensional space, d represents the space dimension (in the pulverized coal pipeline scenario, generally d = 3), σ represents the bandwidth parameter, and X i (t) represents the i-th migration trajectory of the pulverized coal particles in the migration trajectory set at time t.

[0116] By accurately calculating the probability density function, it is possible to clearly understand the probability of pulverized coal particles appearing at different positions in the pipeline. Based on this data, engineers can analyze the flow state of pulverized coal more accurately, providing strong support for optimizing the design and operation of the coal pulverizing system. From the perspective of practical applications, an accurate probability density function helps to improve the efficiency and stability of the coal pulverizing system. In actual production, mastering the distribution probability of pulverized coal particles can reasonably adjust equipment parameters, such as optimizing the grinding intensity of the coal mill and adjusting the wind speed in the pipeline, making the distribution of pulverized coal in the pipeline more uniform, reducing the situation of too high or too low local concentration, thereby effectively improving the combustion efficiency, reducing energy consumption, and at the same time reducing the probability of equipment wear and failure, ensuring the reliable operation of the coal pulverizing system.

[0117] In one of the embodiments, as Figure 2 shown, dynamically adjusting the control instructions for the coal pulverizing process according to the comprehensive monitoring parameters and adaptively calibrating the operating state of the sensor to obtain stable multi-parameter monitoring data of the pulverized coal flow, which may include the following steps:

[0118] Step S701, obtaining the fluctuation value of the pulverized coal flow rate in the comprehensive monitoring parameters.

[0119] Step S702, calculating according to the fluctuation value of the pulverized coal flow rate using the correlation model between the pulverized coal concentration and the pressure gradient to obtain the dynamic adjustment coefficient.

[0120] Step S703, inputting the dynamic adjustment coefficient into the coal pulverizing control instruction to obtain real-time control parameters including the valve opening and the mill speed.

[0121] Step S704, calculating the instruction difference value between the real-time control parameter and the original sensor data. If the instruction difference value exceeds the preset threshold, the adaptive compensation model is called to correct the sensor drift amount to obtain the corrected sensor parameter; the adaptive compensation model is trained and generated based on the historical data of the temperature gradient and the pulverized coal concentration.

[0122] Step S705, performing fusion processing on the corrected sensor parameter and the real-time control parameter to obtain stable multi-parameter monitoring data of the pulverized coal flow.

[0123] Specifically, first, extract the pulverized coal flow velocity fluctuation value from the comprehensive monitoring parameters. This flow velocity fluctuation value reflects the change in the flow velocity of pulverized coal in the pipeline and is the key basis for subsequent control and adjustment. Next, according to the correlation model between pulverized coal concentration and pressure gradient, substitute the obtained pulverized coal flow velocity fluctuation value into it for calculation to obtain the dynamic adjustment coefficient. Subsequently, input the obtained dynamic adjustment coefficient into the pulverizing control instruction, and through a series of operations and conversions, generate real-time control parameters including valve opening and mill speed, etc. These real-time control parameters are the direct instructions for adjusting the operation state of the pulverizing system. After obtaining the real-time control parameters, calculate the instruction difference value between them and the original sensor data to judge the accuracy of the sensor data and the stability of the system operation. If the instruction difference value exceeds the preset threshold, it indicates that there may be problems such as sensor drift. At this time, call the adaptive compensation model trained based on the historical data of temperature gradient and pulverized coal concentration to correct the drift amount of the sensor, and then obtain the corrected sensor parameters. Finally, perform fusion processing on the corrected sensor parameters and the real-time control parameters. Through a specific fusion algorithm, make full use of the advantages of both to finally obtain stable multi-parameter monitoring data of pulverized coal flow.

[0124] In this embodiment, by monitoring the pulverized coal flow velocity fluctuation value and calculating the dynamic adjustment coefficient, the pulverizing control instruction can be adjusted in real time according to the pulverized coal flow state, making key operating parameters such as valve opening and mill speed more in line with the actual production requirements, effectively improving the operation efficiency of the pulverizing system and the quality of pulverized coal, and reducing energy waste and equipment loss. The monitoring and correction of sensor drift, as well as parameter fusion processing, ensure the accuracy and stability of the multi-parameter monitoring data of pulverized coal flow. Stable and reliable data provides accurate system operation information for operators, helping to timely discover potential problems and take corresponding measures.

[0125] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are sequentially shown according to the indication of the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily need to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0126] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the refined control method of the coal pulverizing system based on the coal powder flow characteristics as described above are implemented.

[0127] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0128] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial descriptions of the method embodiments. The device embodiments described above are only illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0129] The above embodiments only represent several implementation manners of the embodiments of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A fine control method for a pulverizing system based on the flow characteristics of pulverized coal, characterized in that: The method comprises: Acquire multi-band original acoustic wave signals in coal powder flow; perform spectrum decomposition and feature extraction adjustment on the original acoustic wave signals to obtain real-time change trend of coal powder fineness distribution; Obtaining the charged signal strength of the coal powder particles and analyzing the multi-point electrostatic field strength in combination with the real-time change trend to obtain an accurate measurement value of the coal powder flow rate; The real-time change trend and the precise measurement value are input into a trained particle kinematics model to deduce the mass distribution characteristics of the coal powder in the pipeline, and a dynamic estimation value of the coal powder concentration is calculated; The real-time change trend, the precise measurement value and the dynamic estimation value of the concentration are integrated to generate comprehensive monitoring parameters of the coal powder flow; According to the comprehensive monitoring parameters, the control instructions of the pulverizing process are dynamically adjusted and the operating state of the sensor is adaptively calibrated to obtain stable multi-parameter monitoring data of coal powder flow.

2. The method according to claim 1, characterized in that: The spectrum decomposition and feature extraction adjustment of the original acoustic wave signal are performed to obtain the real-time change trend of the coal powder fineness distribution, including: Performing frequency domain conversion on the original sound wave signal to obtain energy distribution data of each frequency band; the original sound wave signal includes the sound wave characteristics of collision of particles with different particle sizes; According to the mapping relationship between the coal powder particle size and the acoustic wave frequency band, the dynamic clustering algorithm is used to perform feature decoupling on the energy distribution data of each frequency band, and separate the frequency band energy weight coefficient related to the coal powder fineness; Based on the frequency band energy weight coefficient, a frequency band energy change matrix is ​​constructed to obtain the frequency band energy change matrix; the dimension of the frequency band energy change matrix corresponds to the coal powder particle size distribution range; Inputting the frequency band energy variation matrix into the trained coal powder fineness regression model to obtain the particle size cumulative distribution function of the coal powder particles; The real-time distribution curve of coal powder fineness is calculated according to the particle size cumulative distribution function to obtain the real-time change trend of coal powder fineness distribution.

3. The method according to claim 2, characterized in that According to the mapping relationship between the coal powder particle size and the acoustic wave frequency band, the dynamic clustering algorithm is used to perform feature decoupling on the energy distribution data of each frequency band, and separate the frequency band energy weight coefficient related to the coal powder fineness, including: Acquire the mapping relationship data between the coal powder particle size and the sound wave frequency band; the mapping relationship data includes the frequency band division rules corresponding to different particle sizes; Perform frequency band decomposition on the mapping relationship data according to the frequency band division rule to generate an energy distribution matrix for each frequency band; Iteratively calculating the energy distribution matrix using a dynamic clustering algorithm to obtain a sample classification vector; Extracting a characteristic frequency band index value related to coal powder fineness according to the sample attribution category vector; Based on the characteristic frequency band index value, a target frequency band energy sequence is intercepted, and a proportion parameter of the target frequency band energy sequence in the total energy is calculated; A frequency band weight optimization model is established according to the proportion parameters using a linear mapping relationship to obtain a frequency band energy weight coefficient.

4. The method according to claim 1, characterized in that: The acquisition of the charged signal strength of the coal powder particles and the analysis of the multi-point electrostatic field strength in combination with the real-time change trend to obtain an accurate measurement value of the coal powder flow rate include: Generate a time series feature matrix according to the strength of the charged signal; the time series feature matrix includes the electric field fluctuation amplitude and phase difference of each monitoring point; The time series feature matrix is ​​input into a dynamic calibration model to extract the signal noise component to obtain a high-frequency interference pattern; the dynamic calibration model separates the signal noise component based on a sliding window mechanism; The high-frequency interference pattern is spatially matched with the electrostatic field distribution parameter to obtain a flow rate compensation coefficient; the flow rate compensation coefficient is used to correct the initial flow rate value; The velocity compensation coefficient and the initial velocity value are integrated by using a regression algorithm to obtain an accurate measurement value of the pulverized coal velocity.

5. The method according to claim 4, characterized in that The spatial matching of the high-frequency interference pattern with the electrostatic field distribution parameters to obtain the flow rate compensation coefficient includes: Generate grid alignment parameters of electrostatic field distribution parameters according to the frequency feature set of the high-frequency interference mode; the frequency feature set includes interference intensity distribution of multiple spatial discrete points; Interpolating and reconstructing the electrostatic field distribution parameters based on the grid alignment parameters to obtain an electrostatic field gradient matrix after spatial matching; The electrostatic field gradient matrix and the frequency feature set are input into a compensation model to obtain a flow rate compensation coefficient carrying a weight factor.

6. The method according to claim 1, characterized in that The real-time change trend and the accurate measurement value are input into the trained particle kinematics model to deduce the mass distribution characteristics of the pulverized coal in the pipeline, and the dynamic estimated value of the pulverized coal concentration is calculated, including: Acquiring flow velocity fluctuation data and pressure gradient data in the real-time change trend; Generate a flow velocity distribution matrix of the pipeline cross section according to the flow velocity fluctuation data; According to the coupling relationship between the pressure gradient data and the velocity distribution matrix, a set of migration trajectories of the coal powder particles is calculated using a multi-physics field coupling algorithm; Inputting the migration trajectory set into a trained particle kinematics model to obtain a probability density function of the migration trajectory; Performing three-dimensional gridding numerical integration according to the probability density function to obtain the spatial discreteness distribution of coal powder mass; Extracting the mass ratio coefficient of each grid unit from the spatial discreteness distribution, and generating a dynamic concentration gradient map in combination with the vector components of the velocity distribution matrix; Wherein, if there is a sudden change in the mass proportion coefficient in a local area of ​​the dynamic concentration gradient map, the weight parameter of the probability density function is updated to obtain an updated weight parameter; Recalculate the migration trajectory set according to the updated weight parameters to generate a revised dynamic concentration gradient map; The corrected dynamic concentration gradient map is compared with the pressure gradient data for residuals. If the residual value exceeds a preset threshold, the boundary condition parameters of the multi-physical field coupling algorithm are adjusted according to the residual comparison result to obtain a dynamic estimate of the final coal powder concentration.

7. The method according to claim 6, characterized in that The probability density function is calculated by the following formula: Where f(x) represents the probability density function of the migration trajectory, h represents the bandwidth, n represents the number of trajectories contained in the migration trajectory set, K(x,y) represents the kernel function, x, y represent vectors in d-dimensional space, d represents the spatial dimension (in the coal powder pipeline scenario, generally d = 3), σ represents the bandwidth parameter, and X i (t) represents the i-th migration trajectory of the coal powder particle in the migration trajectory set at time t.

8. The method according to claim 1, characterized in that: The control instructions of the pulverizing process are dynamically adjusted according to the comprehensive monitoring parameters and the operating state of the sensor is adaptively calibrated to obtain stable multi-parameter monitoring data of pulverized coal flow, including: Obtaining the coal powder flow rate fluctuation value in the comprehensive monitoring parameters; According to the pulverized coal flow velocity fluctuation value, a correlation model between pulverized coal concentration and pressure gradient is used to calculate to obtain a dynamic adjustment coefficient; The dynamic adjustment coefficient is input into the milling control instruction to obtain real-time control parameters including valve opening and mill speed; The command difference value between the real-time control parameter and the original data of the sensor is calculated. If the command difference value exceeds a preset threshold, the adaptive compensation model is called to correct the sensor drift to obtain the corrected sensor parameter; the adaptive compensation model is generated based on the historical data training of temperature gradient and coal powder concentration; The corrected sensor parameters are fused with the real-time control parameters to obtain stable multi-parameter monitoring data of coal powder flow.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images