Uniform flow field-based staged rotor speed regulation method and system

Abstract

The application relates to the technical field of cement production equipment, in particular to a hierarchical rotor speed regulation method and system based on a uniform flow field, which has the advantages of strong dynamic adaptability, high accuracy, simple operation, strong predictability and strong scalability by introducing the hierarchical rotor speed regulation method based on the uniform flow field, reduces energy consumption and maintenance cost, and improves the working efficiency of cement production grinding operation and the quality of cement products; the method comprises the following steps: obtaining target granularity of grinding operation and cement material granularity distribution data; dividing the cement material granularity distribution data based on the target granularity to obtain cement material grading characteristic information, wherein the cement material grading characteristic information comprises target granularity, target granularity mass proportion, non-target granularity and non-target granularity mass proportion; obtaining hierarchical rotor characteristic information; and fusing the hierarchical rotor characteristic information and the cement material grading characteristic information.

Description

Technical Field

[0001] This invention relates to the technical field of cement production equipment, and in particular to a method and system for regulating rotor speed based on a uniform flow field. Background Technology

[0002] In the cement production process, the goal of grinding is to grind cement materials to a specific particle size to meet different engineering requirements. During the grinding process, the particle size distribution of the material has a decisive influence on the performance of the final product. In order to obtain a uniform particle size distribution, it is necessary to precisely control the speed of the grinding equipment, especially the classifying rotor.

[0003] Most existing methods for controlling the speed of classifying rotors are based on fixed process parameter settings, ignoring the dynamic relationship between material characteristics and the characteristics of the classifying rotor itself during the grinding process. As a result, in practical applications, even under similar process conditions, it is difficult to obtain consistent grinding results. In particular, when processing materials with different particle size distribution characteristics, traditional fixed speed control strategies often cannot meet the requirements for precise control of product particle size distribution. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and system for controlling the speed of a graded rotor based on a uniform flow field. This method has advantages such as strong dynamic adaptability, high precision, simple operation, strong predictability, and strong scalability, thereby reducing energy consumption and maintenance costs and improving the efficiency of cement grinding operations and the quality of cement products.

[0005] In a first aspect, the present invention provides a method for graded rotor speed control based on a uniform flow field, the method comprising:

[0006] Obtain target particle size and cement material particle size distribution data for grinding operations;

[0007] Based on the target particle size, the particle size distribution data of the cement material is divided to obtain the grading characteristic information of the cement material. The grading characteristic information of the cement material includes the target particle size, the mass percentage of the target particle size, the non-target particle size, and the mass percentage of the non-target particle size.

[0008] Obtain the characteristic information of the staged rotor;

[0009] By fusing the characteristic information of the grading rotor and the grading characteristic information of cement materials, a multi-dimensional feature set of grinding operations is obtained.

[0010] The multi-dimensional feature set of the grinding operation is input into the pre-built uniform flow field prediction model to obtain the uniform flow field feature vector of this grinding operation.

[0011] The obtained uniform flow field feature vector is used as an index marker to match the pre-built flow field graded rotor reference space, mapping the optimal graded rotor speed corresponding to the uniform flow field feature vector, and controlling the operation of the graded rotor according to the optimal graded rotor speed.

[0012] Furthermore, the method for obtaining the granularity distribution data includes:

[0013] Select a standard sieve based on the particle size range of the cement material, with the sieve aperture sizes arranged from largest to smallest;

[0014] The cement material is dried, crushed, and pre-screened. Then the cement material is weighed and the initial total weight is recorded.

[0015] Stack the sieves in order of decreasing aperture size, place a lid on top and a receiving tray at the bottom, and evenly spread the cement material on the top sieve. Use mechanical vibration to make the cement material pass through the sieves of the corresponding aperture size until it falls into the bottom tray.

[0016] Vibration stops after all cement material has passed through a permeable sieve;

[0017] After the vibration stops, weigh the residue on each layer of the sieve separately;

[0018] Compare the weight of the material on each sieve with the initial total weight to calculate the percentage of material in each particle size range;

[0019] Plot the particle size distribution curve or cumulative particle size distribution curve using the sieve aperture size as the horizontal axis and the percentage of material passing through each sieve aperture as the vertical axis.

[0020] Furthermore, the graded rotor characteristic information includes rotor size, rotor mass, blade angle, number of blades, blade material, and inlet / outlet diameter.

[0021] Furthermore, the method for data fusion of the stage rotor characteristic information and cement material grading characteristic information includes:

[0022] Remove data points for grading rotor characteristic information and cement material grading characteristic information, and handle missing values;

[0023] Identify features relevant to the fusion target and exclude irrelevant features;

[0024] Select the fusion algorithm based on the data type and fusion objective;

[0025] By applying a fusion algorithm, the data is merged into a unified dataset;

[0026] Verify the accuracy of the fusion results and check for any unreasonable data points.

[0027] Furthermore, the method for constructing the multi-dimensional feature set includes:

[0028] Verify the obtained data on the grading characteristics of cement materials and the characteristics of the grading rotor;

[0029] The data on the grading characteristics of cement materials and the characteristics of the grading rotor are standardized.

[0030] Based on the actual needs of grinding operations and existing industry experience, the characteristic information that has the greatest impact on grinding effect is selected as the key feature.

[0031] Arrange the selected features in order to form a feature vector;

[0032] Data fusion algorithms are used to fuse individual feature vectors into a multi-dimensional feature set.

[0033] Furthermore, the method for constructing the uniform flow field prediction model for grinding includes:

[0034] Collect records of past grinding operations and record environmental variables related to the grinding process;

[0035] Remove outlier data and handle missing data; transform data of different dimensions and scales to the same scale.

[0036] The correlation between features and the target variable is analyzed using the coefficient matrix;

[0037] The feature dimensionality is reduced and redundant features are removed by recursive feature elimination and principal component analysis.

[0038] Select candidate evaluation models based on the nature of the problem;

[0039] The dataset was divided into training, validation and test sets. K-fold cross-validation was used to evaluate the generalization ability of the uniform flow field prediction model for grinding. Grid search and random search methods were used to adjust the parameters of the uniform flow field prediction model for grinding.

[0040] The milling uniform flow field prediction model was trained on the training set using the optimized parameters, and its performance was evaluated on the validation and test sets.

[0041] Based on the evaluation results, adjustments were made to the selection stage of the uniform flow field prediction model for grinding to ensure that the performance of the uniform flow field prediction model for grinding meets the preset standards.

[0042] Furthermore, the method for constructing the flow field staged rotor reference space is as follows:

[0043] Detailed operational data is collected from the grinding operation. The collected operational data is cleaned to remove outliers and fill in missing values. The operational data is then processed according to standards.

[0044] Based on theoretical and practical experience in grinding operations, we selected the features that have the greatest impact on grinding results, tried different combinations of features, and determined the feature combination that best represents the state and effect of grinding operations through correlation analysis and statistical testing.

[0045] Create a feature vector for each historical job instance based on the selected features;

[0046] For each historical operation instance, the grinding effect is evaluated, and based on the evaluation results, the optimal stage rotor speed corresponding to each feature vector is determined;

[0047] The feature vectors are paired with the corresponding optimal graded rotor speeds to form data pairs; a queryable reference space is constructed using data mining or machine learning techniques.

[0048] Using independent validation set data, we examined whether the staged rotor speed predicted by the control space could actually bring about the expected grinding effect. Based on the validation results, we adjusted the feature selection, grinding uniform flow field prediction model parameters, and optimization strategies.

[0049] By integrating the control space into the grinding operation control system, the speed of the grading rotor can be automatically adjusted based on the feature vectors collected in real time.

[0050] On the other hand, this application also provides a staged rotor speed control system based on a uniform flow field, the system comprising:

[0051] The material property analysis module is used to obtain the target particle size of the grinding operation and the original particle size distribution data of cement materials; analyze and extract the grading characteristic information of cement materials, including target particle size, mass percentage of target particle size, mass percentage of non-target particle size and non-target particle size.

[0052] The rotor characteristic reading module is used to acquire the physical characteristic information of the graded rotor, which includes rotor size, rotor mass, blade angle, number of blades, blade material, and inlet / outlet diameter.

[0053] The data fusion module is used to integrate the cement material grading characteristic information output by the material characteristic analysis module with the rotor characteristic information provided by the rotor characteristic reading module to form a multi-dimensional feature set of grinding operation.

[0054] The flow field prediction module receives the multi-dimensional feature set output by the data fusion module; and uses the pre-trained uniform flow field prediction model to calculate the uniform flow field feature vector for this grinding operation.

[0055] The rotor speed optimization module is used to use the uniform flow field feature vector output by the flow field prediction module as a query condition in the flow field graded rotor comparison space; execute the matching algorithm to find the rotor speed setting that best matches the feature vector; and output the optimal graded rotor speed to the execution unit to control the graded rotor to run at that speed.

[0056] A control and execution module is used to adjust the actual operating speed of the staged rotor according to the optimal staged rotor speed command provided by the rotor speed optimization module. Thirdly, this application provides an electronic device including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus, and the computer program, when executed by the processor, implements the steps of any of the methods described above.

[0057] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0058] Compared with existing technologies, the advantages of this invention are as follows: By real-time monitoring and analysis of material characteristics and the physical characteristics of the classifying rotor, the speed of the classifying rotor can be automatically adjusted under constantly changing operating conditions, improving the adaptability to different material characteristics and achieving more flexible grinding operation control; by combining material particle size distribution data with the characteristics of the classifying rotor, the optimal classifying rotor speed is calculated through a predictive model, facilitating precise control of the material particle size distribution and ensuring that the product quality meets engineering requirements; by adopting a dynamic adjustment strategy based on material characteristics, a stable grinding effect is maintained when processing materials with different particle size distribution characteristics, improving product quality and consistency; and by optimizing the speed of the classifying rotor... This reduces energy consumption and the occurrence of over- or under-grinding, lowering production costs and energy consumption. Through flow field prediction models and data fusion technology, it enhances the intelligence of grinding operations, reduces reliance on manual labor, and increases the automation level of the production process. Based on dynamic analysis of material and rotor characteristic data, it adjusts and optimizes according to changes in new material types and production demands, exhibiting good scalability and adaptability. By introducing a staged rotor speed control method based on a uniform flow field, it offers advantages such as strong dynamic adaptability, high accuracy, simple operation, strong predictability, and strong scalability, reducing energy consumption and maintenance costs, and improving the efficiency of cement grinding operations and the quality of cement products. Attached Figure Description

[0059] Figure 1 This is a flowchart of the present invention;

[0060] Figure 2 This is a flowchart of the method for constructing a prediction model for a uniform flow field in grinding.

[0061] Figure 3 This is a structural diagram of a staged rotor speed control system based on a uniform flow field. Detailed Implementation

[0062] As will be apparent to those skilled in the art from the description of this application, this application can be implemented as a method, apparatus, electronic device, and computer-readable storage medium. Therefore, this application can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. Furthermore, in some embodiments, this application can also be implemented as a computer program product contained in one or more computer-readable storage media, which includes computer program code.

[0063] The aforementioned computer-readable storage medium may be any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory, flash memory, optical fiber, optical disc read-only memory, optical storage devices, magnetic storage devices, or any combination thereof. In this application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0064] The acquisition, storage, use, and processing of data in this application all comply with relevant legal provisions.

[0065] This application describes the provided methods, apparatus, and electronic devices using flowcharts and / or block diagrams.

[0066] It should be understood that each block of a flowchart and / or block diagram, as well as combinations of blocks in a flowchart and / or block diagram, can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine that, when executed by a computer or other programmable data processing apparatus, creates means for implementing the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0067] These computer-readable program instructions may also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to function in a particular manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction apparatus product that includes the functions / operations specified in the blocks of a flowchart and / or block diagram.

[0068] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer-implemented process, such that the instructions that execute on the computer or other programmable data processing apparatus provide a process for implementing the functions / operations specified in the blocks of the flowchart and / or block diagram.

[0069] This application will now be described with reference to the accompanying drawings.

[0070] Example 1: As Figures 1 to 2 As shown, the stepwise rotor speed control method based on a uniform flow field of the present invention specifically includes the following steps:

[0071] S1. Obtain the target particle size and particle size distribution data of cement materials for grinding operations;

[0072] The specific implementation steps for S1 are as follows:

[0073] The target particle size for grinding operations is set based on the final use and performance requirements of the cement. In cement production, the setting of the target particle size should take into account the following factors:

[0074] First, analyze the final use and performance requirements of cement. Different projects have different requirements for cement's hydration rate, early strength, final strength, fluidity, and durability. These properties are directly related to the cement's particle size distribution. For projects requiring rapid hardening, finer cement particles are needed to accelerate the hydration process. For concrete preparation requiring good fluidity, it is necessary to control the cement particle size distribution to avoid excessive fineness affecting fluidity. Refer to national or international standards, such as ASTM and EN, which have clear regulations on cement particle size distribution. Consult relevant standards to determine a reasonable range for the target particle size. Industry specifications and internal company standards also provide specific particle size requirements. Investigate market demand for cement products, especially the preferences for cement particle size in different application areas. Special customer requirements, such as ultrafine cement and low-alkali cement, will directly affect the setting of the target particle size. Through experimental research, evaluate the impact of different particle size distributions on cement performance. Then, use historical production data to analyze successful particle size distribution cases in similar products or applications as a reference for setting the target particle size. Based on the above information, determine the target particle size for grinding operations to improve the applicability of cement products.

[0075] Cement particle size distribution data describes the distribution of cement particle sizes before it enters the grinding equipment. This data directly reflects the original particle size state of the material, facilitating the evaluation of grinding operations and the adjustment of grinding parameters. The data is presented as a particle size distribution curve or histogram, where the horizontal axis represents particle size and the vertical axis represents the mass or volume ratio of particles within the corresponding size range. The methods for obtaining particle size distribution data are as follows:

[0076] Screening method: Select standard sieves according to the particle size range of cement materials, arranging the sieve apertures from largest to smallest; dry the cement materials and perform crushing and pre-screening; weigh a portion of the cement materials and record the initial total weight; stack the sieves in order of aperture size from largest to smallest, placing a lid on top and a receiving tray at the bottom; evenly spread the cement materials on the top layer of sieves, and use mechanical vibration to make the cement materials pass through the sieves of the corresponding aperture size and fall into the next layer of sieves or the tray, screening until all cement materials have passed through a penetrable sieve or remain on a certain layer of sieves without moving; after stopping vibration, weigh the residue on each layer of sieves; compare the weight of the material on each layer of sieves with the initial total weight to calculate the percentage of material in each particle size range; plot the particle size distribution curve or cumulative particle size distribution curve with the sieve aperture size as the x-axis and the percentage of material passing through each sieve aperture as the y-axis;

[0077] Laser particle size analysis: Utilizing the principle of laser scattering, the particle size distribution of materials is calculated by measuring the intensity distribution of the scattered light;

[0078] Image analysis method: Take particle images using a microscope or scanning electron microscope, and then use image processing software to analyze particle size distribution;

[0079] Sedimentation method: suitable for fine particles, it determines particle size distribution by measuring the settling velocity of particles in a liquid;

[0080] The final application of cement is determined by comprehensively considering its end use, performance requirements, market trends, specific customer requirements, and industry standards, ensuring the rationality and applicability of the target particle size. National or international standards are referenced to ensure cement product quality meets industry specifications and enhances market competitiveness. Experimental studies evaluate the impact of different particle size distributions on cement performance, ensuring the scientific validity and feasibility of the target particle size setting. Historical production data and successful case studies are analyzed to provide empirical evidence for setting the target particle size, reducing trial-and-error costs. The target particle size is set flexibly to adapt to the varying needs of different engineering applications and changing market and customer demands.

[0081] Using sieving, laser particle size analysis, image analysis, and sedimentation methods, accurate particle size distribution data can be provided, ensuring precise control of grinding operations. Different particle size analysis methods are applicable to materials with different particle size ranges, ensuring comprehensive data acquisition covering materials from coarse to extremely fine particles. The application of modern technologies such as laser particle size analysis and image analysis improves the efficiency and accuracy of particle size distribution data acquisition. Accurate particle size distribution data provides a data foundation for monitoring and quality control of grinding operations, facilitating timely adjustment of process parameters and improving product quality stability.

[0082] S2. Divide the cement material particle size distribution data based on the target particle size to obtain cement material grading characteristic information. The cement material grading characteristic information includes target particle size, target particle size mass ratio, non-target particle size, and non-target particle size mass ratio.

[0083] The specific implementation steps for S2 are as follows:

[0084] First, the particle size distribution data of cement materials is divided based on the target particle size. The target particle size for grinding operations is clearly defined, and the original particle size distribution data of cement materials is obtained through experimental methods. The particle size distribution data is displayed in the form of a particle size distribution curve or histogram, where the horizontal axis represents particle size and the vertical axis represents the material mass or volume percentage for each particle size segment. The original particle size distribution data is divided according to the target particle size. The particle size distribution curve is divided into two parts: "target particle size" and "non-target particle size." The "target particle size" part includes all particle size ranges falling within the target particle size range; the "non-target particle size" part... This covers all particle size ranges outside the target particle size range, including coarse particles larger than the target particle size and excessively fine particles smaller than the target particle size; it calculates the mass percentage of material within the target particle size range, i.e., the percentage of material smaller than or equal to the target particle size in the total material mass; then it calculates the mass percentage of material outside the target particle size range, i.e., the percentage of material outside the target particle size range in the total material mass; it analyzes the mass percentages of the target particle size range and the non-target particle size range to assess the classification characteristics of the material; it facilitates understanding the particle size state of the material before the grinding operation, and the degree of improvement that needs to be achieved through the grinding operation;

[0085] Next, the grading characteristics information of cement materials was extracted. The extracted grading characteristics information of cement materials includes the following:

[0086] Target particle size: Target particle size refers to the particle size standard that the grinding operation aims to achieve, which is set according to the end use and performance requirements of cement;

[0087] Target particle size mass percentage: refers to the percentage of the total mass of cement material that meets the target particle size requirements;

[0088] Non-target particle size: Non-target particle size refers to materials whose particle size is larger or smaller than the target particle size range;

[0089] Non-target particle size mass percentage: Non-target particle size mass percentage refers to the percentage of non-target particle size material in the total material mass;

[0090] By refining particle size distribution data into target and non-target particle sizes, this method achieves refined management of material particle size distribution, facilitating precise control of particle size output during grinding operations. Setting the target particle size provides clear directionality for grinding operations, ensuring the production process focuses on achieving predetermined particle size standards, thus improving production targeting and efficiency. Calculating and analyzing the mass ratio of target and non-target particle sizes provides data support for adjusting grinding parameters. A mechanism for real-time monitoring of material particle size distribution changes enables production managers to quickly identify and respond to problems arising during grinding. This reduces over-grinding or under-grinding, lowering unnecessary energy consumption and material loss, and improving resource utilization efficiency and economic benefits. Precise control of material particle size distribution ensures the quality stability and consistency of cement products, meeting the specific needs of different engineering projects. In summary, this step improves grinding efficiency and cement product quality, promotes the rational use of resources and environmental protection, and enhances refined management and technological innovation in cement production.

[0091] S3. Obtain the staged rotor characteristic information, which includes rotor size, rotor mass, blade angle, number of blades, blade material, and inlet / outlet diameter.

[0092] The characteristic information of the graded rotor is as follows:

[0093] Rotor dimensions: The diameter and length of the rotor are measured using measuring tools, and the data are recorded in detail. The diameter and length of the rotor determine the rotor's processing capacity and grinding efficiency. A larger diameter generally represents a higher processing capacity, while the rotor length affects the residence time of materials in the classification zone and the classification effect.

[0094] Rotor mass: The rotor mass is calculated based on the rotor's dimensions and material density, or the accurate mass is obtained through actual weighing. The rotor mass affects the rotor's dynamic balance and dynamic stability during high-speed rotation. The greater the rotor mass, the higher the dynamic balance and dynamic stability of the rotor during high-speed rotation.

[0095] Blade angle: The blade angle is measured using an angle measuring tool. The blade angle design changes the distribution of airflow inside the rotor and affects the material classification effect. A suitable blade angle helps to achieve a uniform material particle size distribution.

[0096] Number of blades: The number of blades is obtained visually; more blades can provide finer grading, but too many blades will increase air resistance, reduce airflow speed, and affect the grading effect.

[0097] Blade material: The blade material can be obtained by checking the rotor's material list; different materials have different wear resistance, density and hardness. Selecting the right material can extend the blade's service life and reduce maintenance costs.

[0098] Inlet and outlet diameters: Use measuring tools to measure and record the diameters of the inlet and outlet; the size of the inlet and outlet diameters directly affects the inlet and outlet speeds of materials and airflow, thus affecting the classification efficiency; appropriate inlet and outlet diameter design can optimize the material throughput and airflow distribution;

[0099] By meticulously recording rotor dimensions, mass, blade angles, quantity, materials, and inlet / outlet diameters, refined management of the classifying rotor is achieved. This facilitates precise rotor speed control, ensuring materials are classified to the expected particle size. Information on rotor dimensions and mass allows for optimization of rotor capacity and dynamic balance, improving stability and efficiency at high speeds. Optimization of blade angles and quantity improves airflow distribution, enhances classification accuracy, reduces airflow resistance caused by excessive blades, and maintains high-efficiency classification. Understanding blade material characteristics allows for the selection of highly wear-resistant materials, extending rotor lifespan, reducing maintenance frequency and costs, and improving production continuity and economy. Reasonable inlet / outlet diameter design optimizes material throughput and airflow distribution, reducing unnecessary energy consumption and improving grinding efficiency. Understanding the comprehensive characteristics of the classifying rotor allows for flexible adjustment of rotor parameters to meet different material characteristics and particle size distribution requirements, enhancing the adaptability and flexibility of grinding operations. Optimization of rotor mass and dynamic balance reduces the risk of equipment failure and unexpected downtime, improving production safety and reliability. Comprehensive collection and analysis of classifying rotor characteristic information provides a foundation for optimizing grinding operations, reducing resource waste, and improving production efficiency and product quality.

[0100] S4. Data fusion of the graded rotor characteristic information and cement material graded characteristic information to obtain a multi-dimensional feature set of grinding operation;

[0101] The specific implementation steps for S4 are as follows:

[0102] The grading rotor characteristic information and cement material grading characteristic information are integrated using the following method:

[0103] Data preprocessing: Remove invalid or erroneous data points from the grading rotor characteristic information and cement material grading characteristic information, and handle missing values; transform the data to the same scale to avoid differences in units or numerical ranges affecting subsequent processing; convert the data into a form suitable for fusion as needed;

[0104] Feature extraction: Identify the most important features for the fusion target and eliminate redundant or irrelevant features; extract key information from the original data and use dimensionality reduction techniques such as principal component analysis or independent component analysis.

[0105] Fusion algorithm selection: Select an appropriate fusion algorithm based on the data type and fusion objective. Common fusion algorithms include weighted average, Bayesian fusion, neural network fusion, and fuzzy logic fusion.

[0106] Data fusion implementation: Apply the selected fusion algorithm to merge data from different sources into a unified dataset or feature vector; during the fusion process, it is necessary to assign weights to the features of different data sources;

[0107] Results verification and optimization: Verify the accuracy and consistency of the fusion results, and check for any abnormal or unreasonable data points; adjust the parameters or feature selection of the fusion algorithm based on the verification results to optimize the fusion process;

[0108] Application of fusion results: The fused data is used in downstream tasks;

[0109] Data fusion integrates material and equipment characteristics to reflect a multi-dimensional set of features of the grinding operation status; this facilitates improved prediction accuracy and grinding effect control in subsequent uniform flow field prediction.

[0110] The specific method for constructing a multi-dimensional feature set is as follows:

[0111] Data preparation: First, verify the obtained information on the grading characteristics of cement materials and the characteristics of the grading rotor. The above data must be accurate.

[0112] Data standardization: Standardize or normalize the grading characteristics of cement materials and the characteristics of grading rotors to facilitate fair comparison and combination of different types of characteristic information in the subsequent data fusion process;

[0113] Feature selection and weight determination: Based on the actual needs of grinding operations and existing industry experience, the characteristic information that has the greatest impact on grinding effect is selected as the key feature; and the weight of each feature is determined to reflect the importance of the feature in the grinding operation.

[0114] Constructing a feature vector: Arrange the selected features in a certain order to form a feature vector, which contains all the key characteristic information of the grinding operation;

[0115] Feature vector fusion: Using a data fusion algorithm, the various feature vectors are fused into a multi-dimensional feature set; the multi-dimensional feature set comprehensively describes the characteristics of the grinding operation, providing complete input information for subsequent uniform flow field prediction;

[0116] By integrating data fusion with material and equipment characteristics, a multi-dimensional feature set comprehensively reflects the grinding operation status. This provides input data for the grinding uniform flow field prediction model. Standardization ensures fair comparison of different features within the model, improving prediction accuracy and reliability. Based on this multi-dimensional feature set, the grinding operation dynamically adjusts the staged rotor speed when processing materials with different properties, facilitating more precise material classification control and improving grinding efficiency and product quality. Result verification and optimization during data fusion help detect abnormal or unreasonable data points, allowing for timely adjustments to the fusion algorithm or feature selection, and optimizing the grinding operation control strategy. By rationally selecting features and optimizing the fusion algorithm, redundant data processing is reduced, improving data processing efficiency and minimizing cost waste caused by ineffective operations. Through data fusion and the construction of multi-dimensional feature sets, the control precision and production efficiency of grinding operations are improved, supporting continuous optimization and technological innovation in the cement production process.

[0117] S5. Input the multi-dimensional feature set of the grinding operation into the pre-built uniform flow field prediction model to obtain the uniform flow field feature vector of this grinding operation.

[0118] Step S5 is implemented as follows:

[0119] First, a prediction model for the uniform flow field in grinding is constructed, as follows:

[0120] Collect detailed records of past grinding operations, including material particle size distribution, classifying rotor operating parameters, equipment characteristic information, and operating efficiency and energy consumption data; record environmental variables related to the grinding process, such as temperature, humidity, and operating status parameters of the grinding system;

[0121] Remove outlier data, process missing data, and ensure data quality; transform data of different dimensions and scales to the same scale to facilitate processing by the grinding uniform flow field prediction model; select or construct features that are helpful for prediction based on professional knowledge.

[0122] The correlation between features and target variables is analyzed using methods such as correlation coefficient matrix and mutual information; the feature dimensionality is reduced and redundant features are removed by methods such as recursive feature elimination and principal component analysis, thereby improving the efficiency of the grinding uniform flow field prediction model.

[0123] Based on the nature of the problem, select candidate evaluation models; determine the performance metrics for evaluating the models.

[0124] The dataset was divided into training, validation, and test sets. K-fold cross-validation and other methods were used to evaluate the generalization ability of the grinding uniform flow field prediction model. Grid search and random search methods were used to adjust the parameters of the grinding uniform flow field prediction model and find the optimal parameter combination.

[0125] The milling uniform flow field prediction model was trained on the training set using the optimized parameters; and the performance of the milling uniform flow field prediction model was evaluated on the validation and test sets.

[0126] Based on the evaluation results, return to the feature engineering or grinding uniform flow field prediction model selection stage for adjustment; continuously adjust and optimize until the performance of the grinding uniform flow field prediction model meets the preset standards.

[0127] The multi-dimensional feature set is then checked and preprocessed. This preprocessed feature set is used as input data to the grinding uniform flow field prediction model. Upon receiving the input feature set, the model performs calculations based on its internal algorithms and parameters. After calculation, the model outputs a uniform flow field feature vector, which describes the flow state of the material in the staged rotor under the current grinding conditions. This vector reflects the flow of the material in the staged rotor, including parameters such as velocity, flow rate, and pressure distribution. By comparing and analyzing the uniform flow field feature vectors under different conditions, the working performance and efficiency of the staged rotor are evaluated. This evaluation facilitates the control of the staged rotor speed.

[0128] By collecting and analyzing detailed historical grinding operation data, combined with environmental variables and equipment status parameters, the grinding uniform flow field prediction model is built on a sufficient data foundation, improving the comprehensiveness and accuracy of predictions. Outlier removal, missing value handling, and data standardization in the data preprocessing stage ensure the accuracy of the grinding uniform flow field prediction model training and avoid prediction biases caused by data quality issues. Correlation analysis and feature selection techniques reduce noise and redundant features, improving the efficiency of the grinding uniform flow field prediction model and its effective capture of key features. Through multi-feature selection, feature dimensionality reduction, and parameter optimization, the grinding uniform flow field prediction model is continuously optimized and adjusted, improving its generalization ability and prediction accuracy, making it easier to adapt to different grinding operation conditions. Through continuous evaluation and... The uniform flow field prediction model for grinding was adjusted, and a closed-loop system was established from the construction to optimization of the uniform flow field prediction model to ensure continuous improvement and adaptability of the model. The output uniform flow field feature vector provides an intuitive and quantitative evaluation index for grinding operations, facilitating an intuitive understanding of the working state of the classifying rotor and the characteristics of the material flow field, and providing a data foundation for precise control of the classifying rotor speed. Through refined prediction and control, the speed of the classifying rotor can be dynamically adjusted to achieve more efficient material classification, reduce energy consumption, and improve product quality and production efficiency. By combining data-driven approaches with the prediction of the uniform flow field model for grinding, multi-dimensional feature fusion, model optimization and adjustment, and precise control and evaluation, precise control of the speed of the classifying rotor in the cement production process was achieved, improving product quality and performance stability.

[0129] S6. The obtained uniform flow field feature vector is used as an index marker to match the pre-built flow field graded rotor comparison space, mapping the optimal graded rotor speed corresponding to the uniform flow field feature vector, and controlling the operation of the graded rotor according to the optimal graded rotor speed.

[0130] The implementation method for step S6 is as follows:

[0131] First, a flow field staged rotor comparison space is constructed, specifically as follows:

[0132] Detailed operational data are collected from the grinding operation, including various operating speeds of the classifying rotor, detailed equipment parameters, material characteristics, environmental parameters, and performance indicators of the grinding operation; the collected operational data is cleaned to remove outliers and fill in missing values; the data is standardized or normalized to ensure the comparability of data from different sources.

[0133] Based on theoretical and practical experience in grinding operations, we selected the features that have the greatest impact on grinding effect, including flow field characteristics, equipment characteristics, and material characteristics, and constructed a feature set. We tried different combinations of features and determined the feature combination that best represents the state and effect of grinding operations through correlation analysis and statistical testing.

[0134] Based on the selected features, a feature vector is created for each historical job instance. The historical job feature vector contains key information reflecting the flow field state and equipment operating status.

[0135] For each historical operation instance, evaluate its grinding effect; based on the evaluation results, determine the optimal stage rotor speed corresponding to each feature vector;

[0136] The feature vectors are paired with the corresponding optimal stage rotor speeds to form data pairs; a queryable reference space is constructed using data mining or machine learning techniques; the reference space should be able to quickly find the corresponding optimal speed based on the input feature vectors.

[0137] Using independent validation set data, we examined whether the staged rotor speed predicted by the control space could actually bring about the expected grinding effect. Based on the validation results, we adjusted the feature selection, grinding uniform flow field prediction model parameters and optimization strategies until the prediction performance of the control space reached a satisfactory level.

[0138] The control space is integrated into the control system of the grinding operation to realize the automatic adjustment of the stage rotor speed based on the feature vectors collected in real time; as more operating data is accumulated, the control space is continuously updated and optimized.

[0139] Then, the uniform flow field feature vector is used as a query condition and matched in the constructed flow field classification rotor reference space. After matching the corresponding flow field feature, the optimal classification rotor speed is extracted from the reference space. The optimal classification rotor speed value is predicted based on historical experience and the uniform flow field prediction model for grinding, and is considered to achieve the best material classification effect under this specific flow field condition. According to the mapped optimal classification rotor speed, the actual running speed of the classification rotor is adjusted by the automated control system. The control process needs to respond in real time to ensure that the classification rotor speed can be quickly adjusted to the theoretical optimal value to adapt to the changes in material properties during the grinding process. By monitoring the real-time effect of the grinding operation and comparing it with the expected particle size distribution, the classification rotor speed is fine-tuned according to the deviation to achieve closed-loop control and optimize the grinding effect.

[0140] By constructing a flow field staged rotor reference space based on historical operation data, the optimal staged rotor speed can be matched for each specific flow field feature vector, ensuring a highly personalized and dynamically adaptable control strategy when facing different material properties and operating conditions. Based on the collection, cleaning, standardization, and feature selection of large amounts of data, the scientific nature and accuracy of control decisions are ensured, reducing the limitations of traditional fixed parameter settings and improving the efficiency and product quality of grinding operations. The performance of the grinding uniform flow field prediction model is validated using validation set data, and feature selection, grinding uniform flow field prediction model parameters, and optimization strategies are continuously adjusted based on feedback results, ensuring the continuous improvement of the prediction performance of the reference space and forming a closed-loop system of continuous learning and optimization. The integrated control system automatically adjusts the speed of the staged rotor based on real-time acquired feature vectors, enabling rapid response and real-time control, reducing delays and misjudgments caused by human intervention, and improving production efficiency. By monitoring the grinding operation in real time and fine-tuning the deviation between the actual particle size distribution and the expected target, closed-loop control is achieved, ensuring the stability and consistency of the grinding process. It comprehensively considers multiple factors such as flow field characteristics, equipment characteristics, and material characteristics, and through the optimization of feature combinations, fully captures key variables affecting the grinding effect, improving the overall effectiveness of the control strategy. Through highly data-driven and intelligent methods, it improves the accuracy and flexibility of staged rotor speed control in grinding operations, as well as the quality and efficiency control in the cement production process.

[0141] Example 2: Figure 3 As shown, the graded rotor speed control system based on a uniform flow field of the present invention specifically includes the following modules;

[0142] The material property analysis module is used to obtain the target particle size of the grinding operation and the original particle size distribution data of cement materials; analyze and extract the grading characteristic information of cement materials, including target particle size, mass percentage of target particle size, mass percentage of non-target particle size and non-target particle size.

[0143] The rotor characteristic reading module is used to acquire the physical characteristic information of the graded rotor, which includes rotor size, rotor mass, blade angle, number of blades, blade material, and inlet / outlet diameter.

[0144] The data fusion module is used to integrate the cement material grading characteristic information output by the material characteristic analysis module with the rotor characteristic information provided by the rotor characteristic reading module to form a multi-dimensional feature set of grinding operation.

[0145] The flow field prediction module receives the multi-dimensional feature set output by the data fusion module; and uses the pre-trained uniform flow field prediction model to calculate the uniform flow field feature vector for this grinding operation.

[0146] The rotor speed optimization module is used to use the uniform flow field feature vector output by the flow field prediction module as a query condition in the flow field graded rotor comparison space; execute the matching algorithm to find the rotor speed setting that best matches the feature vector; and output the optimal graded rotor speed to the execution unit to control the graded rotor to run at that speed.

[0147] The control and execution module is used to adjust the actual operating speed of the staged rotor according to the optimal staged rotor speed command provided by the rotor speed optimization module.

[0148] By analyzing material and rotor characteristics in real time, the uniform flow field-based staged rotor speed control system can automatically adjust the speed of the staged rotor according to different materials and operating conditions, dynamically optimizing the grinding process and better adapting to different types of cement materials. Based on material characteristic analysis and rotor characteristic reading, the system can generate a multi-dimensional feature set and calculate a uniform flow field feature vector through a flow field prediction model, improving the accuracy of rotor speed control, facilitating a more uniform particle size distribution, and improving the quality of cement products. By optimizing the rotor speed, the system can ensure consistent grinding results across different grinding batches, reducing product inconsistencies caused by fluctuations in process parameters. It also helps improve production efficiency and resource utilization; precise rotor speed control reduces over-grinding and under-grinding, reducing energy and raw material waste and lowering production costs; the design of the uniform flow field graded rotor speed control system reduces the need for manual intervention and improves the overall intelligence level of the production line; the modular design allows the uniform flow field graded rotor speed control system to be integrated with existing cement production equipment, facilitating later maintenance and upgrades to cope with future technological advancements or changes in production needs; the uniform flow field graded rotor speed control system solves the problems of low efficiency and product quality fluctuations in traditional grinding operations, reduces production costs, and improves grinding efficiency and product quality in cement production.

[0149] The various variations and specific embodiments of the uniform flow field-based graded rotor speed control method in the aforementioned Embodiment 1 are also applicable to the uniform flow field-based graded rotor speed control system of this embodiment. Through the foregoing detailed description of the uniform flow field-based graded rotor speed control method, those skilled in the art can clearly understand the implementation method of the uniform flow field-based graded rotor speed control system of this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0150] In addition, this application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected via the bus. When the computer program is executed by the processor, it implements the various processes of the above-described method embodiment for controlling output data and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0151] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for regulating the speed of a uniform flow field based on a stage rotor, characterized in that, The method comprises: acquiring target particle size of grinding operation and cement material particle size distribution data; dividing the cement material particle size distribution data based on the target particle size to obtain cement material classification characteristic information, the cement material classification characteristic information including target particle size, target particle size mass proportion, non-target particle size and non-target particle size mass proportion; acquiring classification rotor characteristic information; performing data fusion of the classification rotor characteristic information and the cement material classification characteristic information to obtain a multi-dimensional feature set of the grinding operation; inputting the multi-dimensional feature set of the grinding operation into a pre-constructed grinding uniform flow field prediction model to obtain a uniform flow field feature vector of the current grinding operation; taking the obtained uniform flow field feature vector as an index mark to perform matching in a pre-built flow field classification rotor comparison space, mapping out an optimal classification rotor speed corresponding to the uniform flow field feature vector, and controlling the classification rotor to operate according to the optimal classification rotor speed; the method for performing data fusion of the classification rotor characteristic information and the cement material classification characteristic information comprises: removing data points of the classification rotor characteristic information and the cement material classification characteristic information, and processing missing values; identifying features of the fusion target and excluding irrelevant features; selecting a fusion algorithm according to the data type and the fusion target; applying the fusion algorithm to combine the data into a unified data set; verifying the accuracy of the fusion result and checking whether there are unreasonable data points; the method for constructing the grinding uniform flow field prediction model comprises: collecting records of past grinding operations, and recording environment variables related to the grinding process; removing abnormal data and processing missing data, and converting data of different dimensions and scales to the same scale; analyzing the correlation between the characteristics and the target variables through a coefficient matrix; reducing the feature dimension and removing redundant features through recursive feature elimination and principal component analysis methods; selecting a candidate evaluation model according to the nature of the problem; dividing the data set into a training set, a validation set and a test set, using a K-fold cross-validation method to evaluate the generalization ability of the grinding uniform flow field prediction model, using a grid search and a random search method to adjust the parameters of the grinding uniform flow field prediction model; training the grinding uniform flow field prediction model on the training set using the optimized parameters, and evaluating the performance of the grinding uniform flow field prediction model on the validation set and the test set; according to the evaluation result, returning to the grinding uniform flow field prediction model selection stage for adjustment, so that the performance of the grinding uniform flow field prediction model meets the preset standard.

2. The uniform flow field based staging rotor speed regulation method of claim 1, wherein, the method for acquiring the particle size distribution data comprises: selecting a standard sieve according to the particle size range of the cement material, and arranging the sieve hole sizes from large to small; drying the cement material, crushing and pre-screening it, then weighing the cement material and recording the initial total weight; stacking the sieves from large to small in order of aperture, placing a cover on the top and a receiving tray on the bottom, evenly spreading the cement material on the topmost sieve, and using mechanical vibration to make the cement material pass through the sieve with the corresponding aperture until it falls into the bottom tray; stopping the vibration after all the cement material passes through the penetrable sieve; after stopping the vibration, weighing the residues on each layer of sieve respectively; Compare the weight of the material on each layer of sieve with the initial total weight to calculate the percentage of material in each particle size range; Draw a particle size distribution curve or cumulative particle size distribution curve with the sieve size as the horizontal coordinate and the percentage of material passing through each sieve size as the vertical coordinate.

3. The uniform flow field based staging rotor speed regulation method of claim 1, wherein, The classification rotor characteristic information includes rotor size, rotor mass, blade angle, blade number, blade material, and inlet and outlet diameters.

4. The uniform flow field based staging rotor speed regulation method of claim 1, wherein, The method for constructing the multi-dimensional feature set comprises: Check the acquired classification characteristic information of the cement material and the characteristic information of the classification rotor; Standardize the data of the classification characteristic information of the cement material and the characteristic information of the classification rotor; According to the actual requirements of the grinding operation and the existing industry experience, select the characteristic information that has the greatest impact on the grinding effect as the key feature; Arrange the selected features in order to form a feature vector; Use a data fusion algorithm to fuse each feature vector into a multi-dimensional feature set.

5. The uniform flow field staged rotor speed regulation method of claim 1, wherein, The flow field classification rotor control space building method is as follows: Collect detailed operation data from the grinding operation, clean the collected operation data, remove outliers and fill in missing values, and standardize the operation data; Based on the theory and practical experience of the grinding operation, select the features that have the greatest impact on the grinding effect, try different feature combinations, and determine the feature combination that best represents the state and effect of the grinding operation through correlation analysis and statistical testing; According to the selected features, create a feature vector for each historical operation instance; For each historical operation instance, evaluate the grinding effect, and according to the evaluation result, determine the optimal classification rotor speed corresponding to each feature vector; Pair the feature vector with the corresponding optimal classification rotor speed to form a data pair; use data mining or machine learning techniques to build a queryable control space; Use independent validation set data to verify whether the classification rotor speed predicted by the control space can actually bring about the expected grinding effect, and according to the verification result, adjust the feature selection, grinding uniform flow field prediction model parameters and optimization strategy; Integrate the control space into the control system of the grinding operation to automatically adjust the classification rotor speed according to the real-time collected feature vector.

6. A uniform flow field staged rotor speed regulation system based on the method of claim 1, wherein, The system comprises: A material characteristic analysis module for acquiring target particle size of the grinding operation and original particle size distribution data of the cement material; analyzing and extracting classification characteristic information of the cement material, including target particle size, mass percentage of target particle size, non-target particle size and mass percentage of non-target particle size; A rotor characteristic reading module for acquiring classification rotor physical characteristic information, including rotor size, rotor mass, blade angle, blade number, blade material and inlet and outlet diameters; A data fusion module for integrating the classification characteristic information of the cement material output by the material characteristic analysis module and the rotor characteristic information provided by the rotor characteristic reading module to form a multi-dimensional feature set of the grinding operation; A flow field prediction module for receiving the multi-dimensional feature set output by the data fusion module; using a pre-trained grinding uniform flow field prediction model to calculate the uniform flow field feature vector of the current grinding operation; a rotor speed optimization module, configured to use the uniform flow field feature vector outputted by the flow field prediction module as a query condition in a flow field stage rotor vs. space database, execute a matching algorithm to find the rotor speed setting that best matches the feature vector, and output the optimal stage rotor speed to an execution unit to control the stage rotor to run at the speed; a control and execution module, configured to adjust the actual running speed of the stage rotor according to the optimal stage rotor speed instruction provided by the rotor speed optimization module.

7. A uniform flow field based fractional rotor speed regulation electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor, the transceiver, the memory, and the processor connected by the bus, wherein, The computer program, when executed by the processor, implements the steps in the method of any one of claims 1-5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps in the method of any one of claims 1-5.

Images