Method and system for air volume control of a non-powered classifier mill

Abstract

The present application relates to the technical field of powder concentrator control, in particular to a kind of unpowered powder concentrator air volume control method and system, which reduces the time and frequency of shutdown adjustment through real-time and dynamic control, improves the continuity and stability of production line, finally improves the overall production efficiency;Method comprises: collecting several kinds of material classification samples obtained by unpowered powder concentrator classification powder selection;Based on the preset material classification target feature, target deviation feature extraction is carried out on each material classification sample, and the material classification deviation vector corresponding to the acquisition time is obtained;Material classification deviation vector of multiple acquisition times is obtained, and is arranged in time sequence, and material classification deviation matrix is obtained;The material classification deviation matrix is input into the pre-constructed powder concentrator classification accuracy evaluation model, and the material classification accuracy index corresponding to the unpowered powder concentrator is obtained.

Description

Technical Field

[0001] This invention relates to the technical field of air classifier control, and in particular to a method and system for controlling the air volume of a non-powered air classifier. Background Technology

[0002] In the process of material processing and screening, the non-powered air classifier is a common piece of equipment. It achieves the classification and powdering of materials by means of natural wind or mechanical wind without a power source. However, in practical applications, due to the influence of various factors such as material properties, environmental conditions and the equipment itself, the classification and powdering effect of the non-powered air classifier often fails to meet the expected target, resulting in low material classification accuracy, which affects the quality of cement and other products and production efficiency.

[0003] Existing methods for controlling the airflow of non-powered air classifiers often rely on the operator's experience for adjustment, which is not only inefficient but also makes it difficult to guarantee the accuracy and stability of airflow adjustment. At the same time, due to the lack of real-time monitoring and evaluation of the material classification effect, operators often cannot promptly identify and resolve problems in the classification process, thereby further reducing the material classification accuracy and production efficiency. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and system for controlling the airflow of a non-powered air classifier by reducing downtime and adjustment time and frequency through real-time and dynamic control, thereby improving the continuity and stability of the production line and ultimately enhancing overall production efficiency.

[0005] In a first aspect, the present invention provides a method for controlling the air volume of a non-powered air classifier, the method comprising:

[0006] Several material classification samples were collected from the non-powered air classifier.

[0007] Based on the preset material classification target features, target deviation features are extracted for each material classification sample to obtain the material classification deviation vector corresponding to the collection time.

[0008] Obtain material classification deviation vectors from multiple collection times, arrange them in chronological order, and obtain the material classification deviation matrix;

[0009] The material classification deviation matrix is ​​input into the pre-constructed classification accuracy evaluation model of the air classifier to obtain the material classification accuracy index corresponding to the non-powered air classifier.

[0010] Based on the material classification accuracy index, the optimal air volume control parameters corresponding to the non-powered air classifier are mapped from the preset air volume control parameter database.

[0011] The air volume of the non-powered air classifier is controlled according to the optimal air volume control parameters.

[0012] Furthermore, the material classification deviation vector consists of several deviation feature values, each of which corresponds to a material classification sample. The material classification target features include the target particle size and the target particle size mass ratio corresponding to each material classification sample. The deviation feature value refers to the difference between the mass ratio of the target particle size material and the target particle size mass ratio in the material classification sample.

[0013] Furthermore, in the material classification deviation matrix, the deviation feature values ​​collected at different times corresponding to the same target particle size are located in the same column, and the deviation feature values ​​collected at the same time corresponding to different target particle sizes are located in the same row;

[0014] The material classification deviation matrix is ​​as follows:

[0015]

[0016] Where dmn is an element in the material classification deviation matrix, representing the deviation feature value of the nth particle size level at the mth collection time; the mth row contains the deviation feature values ​​of all particle size levels, corresponding to the mth collection time point; the nth column contains the deviation feature values ​​of the nth particle size level at all collection time points.

[0017] Furthermore, the method for collecting several material classification samples obtained from classification using a non-powered air classifier includes:

[0018] Determine and sample the locations of the non-powered air classifier; the sampling locations include the different grade inlets and outlets of the air classifier.

[0019] The sampling frequency and cycle are determined based on production needs and equipment operating conditions;

[0020] Select sampling tools and equipment for sampling;

[0021] The sampling method is selected based on the location of the sampling points and the properties of the material; sampling methods include continuous flow sampling, manual-assisted sampling, and mechanical-assisted sampling.

[0022] After sampling, each collected material graded sample is identified and recorded, including sampling time, sampling point, and sample number information.

[0023] Furthermore, the method for constructing the classifier grading accuracy evaluation model includes:

[0024] Information on materials before and after classification in a non-powered air classifier under different airflow control parameter settings was collected as the training basis for the classification accuracy evaluation model of the air classifier.

[0025] Determine the target particle size range and the particle size distribution curve of the target mass percentage for each particle size segment;

[0026] Features are extracted from the graded materials, including the percentage of materials in different particle size ranges, the recovery ratio of coarse and fine powder, and the recovery rate of fine powder minus the proportion of fine powder in coarse powder.

[0027] Calculate the deviation between the actual particle size distribution and the target particle size distribution;

[0028] A machine learning algorithm was selected to construct a model for evaluating the grading accuracy of the air classifier.

[0029] The material classification deviation matrix and corresponding classification accuracy index from historical data were used as training datasets to train the air classifier classification accuracy evaluation model; and the parameters of the air classifier classification accuracy evaluation model were adjusted.

[0030] The generalization ability of the air classifier classification accuracy evaluation model was evaluated using independent validation and test sets.

[0031] Furthermore, the method for constructing the air volume control parameter database includes:

[0032] List the airflow control parameters that affect the grading effect; design an experimental matrix, which includes the airflow control parameters that affect the grading effect.

[0033] Perform the experiment by changing only one airflow control parameter while keeping the other airflow control parameters unchanged. Repeat the experiment multiple times according to the number of airflow control parameters and record the parameter settings and results of each experiment.

[0034] Check and clean the experimental data, remove outliers and erroneous data points, convert the cleaned experimental data into a uniform format, and store the converted experimental data in the database;

[0035] Design a database for storing and querying experimental data. The database structure includes tables, fields, and data types.

[0036] Use the database management module to create a database of air volume control parameters and import the pre-processed data.

[0037] Furthermore, the air volume control parameters include the air inlet opening, the air outlet opening, the guide vane angle, the air pressure, and the temperature and humidity inside and outside the classifier hopper.

[0038] On the other hand, this application also provides an airflow control system for a non-powered air classifier, the system comprising:

[0039] The sample acquisition module is used to collect various material classification samples during the classification process of the non-powered air classifier;

[0040] The feature extraction module is used to analyze each material grading sample through preset material grading target features, calculate and extract the difference between the actual mass ratio of the target particle size material and the target set value, and generate a material grading deviation vector;

[0041] The deviation matrix generation module is used to collect and organize material classification deviation vectors at multiple time points to form a time series material classification deviation matrix. The columns of the material classification deviation matrix represent the deviation feature values ​​at different time points, and the rows represent the deviation feature values ​​of different target granularities at the same time point.

[0042] The classification accuracy evaluation module receives the deviation matrix as input and uses a pre-built classification accuracy evaluation model of the air classifier to calculate the material classification accuracy index of the non-powered air classifier.

[0043] The parameter mapping module is used to search and map the optimal air volume control parameters that best match the non-powered air classifier based on the calculated material classification accuracy index in the preset air volume control parameter database.

[0044] The execution module receives the optimal airflow control parameters from the airflow control parameter mapping module, dynamically adjusts the airflow of the non-powered air classifier, and optimizes the grading and classification effect.

[0045] 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.

[0046] 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.

[0047] Compared with existing technologies, the beneficial effects of this invention are as follows: By collecting material classification samples, extracting target deviation features, and constructing a material classification deviation matrix, the classification accuracy is evaluated based on a pre-constructed air classifier classification accuracy evaluation model, and the optimal airflow control parameters are mapped to achieve intelligent and automated airflow control, thus improving work efficiency; by collecting and evaluating the effect of material classification in real time, the airflow control parameters are dynamically adjusted according to the evaluation results, ensuring that the non-powered air classifier always works in the best state, improving material classification accuracy and production efficiency; through the preset material classification target features and the air classifier classification accuracy evaluation model, the material classification accuracy index is accurately calculated, based on the material classification accuracy. The material classification accuracy index maps to the optimal airflow control parameters in the database, improving the accuracy and stability of airflow adjustment. By adapting to changes in different material properties, environmental conditions, and equipment status, the airflow control parameters are dynamically adjusted to optimize the classification and powder selection effect, improving the adaptability and flexibility of classification and powder selection. By reducing human intervention, the operational difficulty and error rate are reduced, improving the reliability and stability of the production process. By optimizing the classification and powder selection effect, the material classification accuracy can be ensured to reach the preset target, improving the quality of products such as cement. Through real-time and dynamic control, the downtime and frequency of adjustments are reduced, improving the continuity and stability of the production line, and ultimately improving overall production efficiency. Attached Figure Description

[0048] Figure 1 This is a flowchart of the present invention;

[0049] Figure 2 This is a flowchart illustrating the construction method of the air classifier grading accuracy evaluation model;

[0050] Figure 3 This is a structural diagram of the air volume control system for a non-powered air classifier. Detailed Implementation

[0051] 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.

[0052] 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.

[0053] The acquisition, storage, use, and processing of data in this application all comply with relevant national laws and regulations.

[0054] This application describes the provided methods, apparatus, and electronic devices using flowcharts and / or block diagrams.

[0055] 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.

[0056] 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.

[0057] 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.

[0058] This application will now be described with reference to the accompanying drawings.

[0059] Example 1: As Figures 1 to 2 As shown, the air volume control method for a non-powered air classifier of the present invention specifically includes the following steps:

[0060] S1. Collect several material classification samples obtained by non-powered air classifier classification;

[0061] The following procedures were followed for collecting material grading samples:

[0062] First, determine the sampling location of the non-powered air classifier and take samples; the selection of sampling points can reflect the material classification inside the air classifier, including the different levels of the air classifier's feed inlet and outlet;

[0063] The sampling frequency and cycle are determined based on production needs and equipment operating conditions. The setting of the sampling frequency and cycle can capture the impact of material properties, environmental conditions, and equipment status factors on the grading effect.

[0064] Select appropriate sampling tools and equipment for sampling; the sampling tools and equipment should be able to accurately collect material samples without causing contamination or damage to the materials;

[0065] Choose an appropriate sampling method based on the location of the sampling points and the properties of the material; for example, for materials with good flowability, use continuous flow sampling; for materials with poor flowability, use manual or mechanically assisted sampling.

[0066] After sampling, each collected material graded sample is identified and recorded, including sampling time, sampling point, and sample number information;

[0067] The collected material grading samples are properly processed to facilitate subsequent analysis. Finally, the processed material grading samples are properly stored to avoid contamination or deterioration.

[0068] Material classification samples serve as the data basis for evaluating the classification effect of the non-powered air classifier. By analyzing and comparing different classification samples, we can understand the particle size distribution, mass ratio, and other indicators of the materials in each classification, and determine whether the classification effect has achieved the expected target. The analysis results of the material classification samples guide adjustments in the production process. When it is found that the quality of the material in a certain classification does not meet the requirements, the air volume control parameters and screen aperture of the non-powered air classifier can be adjusted in time to improve the classification effect, improve product quality and production efficiency. By regularly collecting and analyzing material classification samples, problems in the production process can be identified in a timely manner, and corresponding measures can be taken to solve them, ensuring the stability and consistency of product quality.

[0069] S2. Based on the preset material classification target features, target deviation features are extracted for each material classification sample to obtain the material classification deviation vector corresponding to the collection time. The material classification deviation vector consists of several deviation feature values, each of which corresponds to a material classification sample. The material classification target features include the target particle size and the target particle size mass ratio corresponding to each material classification sample. The deviation feature value refers to the difference between the mass ratio of the target particle size material and the target particle size mass ratio in the material classification sample.

[0070] Step S2, based on the preset material grading target features, extracts the target deviation features for each material grading sample to obtain the material grading deviation vector corresponding to the collection time. The steps are as follows: First, clarify the target features of material grading; for materials such as cement, the target features include the target particle size and the target particle size mass ratio corresponding to each material grading sample.

[0071] For each type of collected material classification sample, the deviation between the actual particle size distribution and the target particle size distribution of the material classification sample is calculated; this is achieved by measuring the mass percentage of material in each particle size range in the sample.

[0072] Then, the measured actual mass percentage is compared with the preset target particle size mass percentage, and the difference is calculated, i.e., the deviation characteristic value; the deviation characteristic value represents the difference between the current material classification sample's material mass percentage within the target particle size range and the target value.

[0073] For each type of material classification sample, a corresponding deviation characteristic value will be obtained. The above deviation characteristic values ​​together constitute the material classification deviation vector corresponding to the collection time.

[0074] The material classification deviation vector is an array containing multiple deviation feature values; each deviation feature value corresponds to a material classification sample and represents the deviation between the material mass ratio of the sample within a specific particle size range and the target value.

[0075] The length of the material classification deviation vector depends on the type and quantity of the material classification samples we collect, as well as the number of particle size ranges we are interested in.

[0076] The calculation process of deviation characteristic values ​​needs to be accurate and reliable to ensure that the effect of material classification can be accurately evaluated. After the deviation characteristic values ​​are calculated, they are arranged and recorded in a specific order to facilitate subsequent analysis and processing.

[0077] By collecting material classification samples in real time and calculating their deviation characteristic values, the current material classification effect can be reflected. This allows operators to respond quickly and adjust airflow control parameters, ensuring continuous improvement in material classification accuracy. Simultaneously, the use of precise measurement and calculation methods guarantees the accuracy of deviation characteristic values, improving the accuracy of evaluation results. Based on preset material classification target characteristics and objective data analysis, interference from human factors is avoided, enhancing the objectivity and fairness of evaluation results and accurately reflecting the material classification effect. Considering the influence of different material classification samples and particle size ranges, multiple deviation characteristic values ​​are calculated and constructed into a material classification deviation vector to comprehensively evaluate the material classification effect, providing better adaptability and flexibility. This system provides clear guidance to operators by evaluating the effectiveness of material grading and calculating deviation characteristics, enabling them to understand current problems and areas for improvement. Furthermore, by reflecting the grading effect in real time, it allows operators to proactively adjust airflow control parameters, reducing potential problems and improving production efficiency. Based on data analysis and machine learning technologies, it enables automatic evaluation and optimization of material grading effectiveness. This improves the intelligence and automation of airflow control, reducing the need for manual intervention and enhancing production stability and reliability. By collecting and analyzing data from graded material samples in real time, it accurately evaluates the grading effect and optimizes airflow control parameters, thereby improving grading accuracy and production efficiency.

[0078] S3. Obtain material classification deviation vectors from multiple collection times and arrange them in chronological order to obtain a material classification deviation matrix. In the material classification deviation matrix, deviation feature values ​​corresponding to the same target particle size collected at different times are located in the same column, and deviation feature values ​​corresponding to different target particle sizes collected at the same time are located in the same row.

[0079] In step S2, the material classification deviation vector at a single data collection point is obtained. This deviation vector quantifies the deviation between the actual classification effect and the preset target. However, data from a single time point only reflects the situation at that time and cannot represent the trend or fluctuation of the classification effect over time, nor can it provide sufficient information for long-term performance evaluation or prediction. Therefore, step S3 collects and organizes deviation vectors from multiple time points to construct a material classification deviation matrix, which facilitates further data analysis. The method for constructing the material classification deviation matrix is ​​as follows:

[0080] In step S2, material classification deviation vectors are collected at multiple collection time points. Each vector corresponds to a specific collection time point. The multiple classification deviation vectors contain information about the deviation between the classification effect of materials of different particle size levels and the target setting, i.e., deviation feature values.

[0081] Each column of the material classification deviation matrix represents a specific granularity level; each column contains deviation characteristic values ​​corresponding to that granularity level collected from all data collection time points; this allows for a longitudinal view of the trend of a specific granularity level changing over time.

[0082] Each row of the material classification deviation matrix represents a specific data collection point; each row contains the deviation feature values ​​of all granularity levels at the same time point; this facilitates horizontal comparison of the classification effects of different granularity levels at a certain time point.

[0083] The number of rows in the material classification deviation matrix is ​​equal to the total number of data collection time points, and the number of columns is equal to the number of particle size levels considered. This ensures that the material classification deviation matrix can clearly reflect the overall picture of the classification effect over time, as well as the interrelationships between different particle size levels.

[0084] The rows of the material classification deviation matrix are arranged in chronological order; this ensures that the evolution trend of classification effect over time can be identified when analyzing the material classification deviation matrix; the chronological order facilitates the use of time series analysis techniques, such as sliding window averaging and trend line fitting, to identify potential patterns or anomalies.

[0085] The material classification deviation matrix is ​​as follows:

[0086]

[0087] Where dmn is an element in the material classification deviation matrix, representing the deviation feature value of the nth particle size level at the mth collection time; the mth row contains the deviation feature values ​​of all particle size levels, corresponding to the mth collection time point; the nth column contains the deviation feature values ​​of the nth particle size level at all collection time points.

[0088] By collecting material classification deviation vectors at multiple time points, the material classification deviation matrix comprehensively reflects the trend and fluctuation of classification effectiveness over time, facilitating more accurate and effective subsequent data analysis. Each column of the material classification deviation matrix represents a specific granularity level, showing the trend of material classification effectiveness at that granularity level over time. This facilitates rapid identification of trends in classification effectiveness and allows for corresponding adjustment measures. Each row represents a specific data collection point, enabling comparison of classification effectiveness at different granularity levels at the same time. Horizontal comparison helps identify the interactions and potential problems between different granularity levels, providing a basis for subsequent optimization. The number of rows in the material classification deviation matrix... The number of columns equals the total number of data collection points, and the number of columns equals the number of particle size levels considered. This clear data structure allows operators to intuitively understand the overall picture of the grading effect over time and the interrelationships between different particle size levels. The rows of the material grading deviation matrix are arranged in chronological order, facilitating the application of time series analysis techniques to identify potential patterns or anomalies. This facilitates the timely detection and resolution of problems in the grading and powder selection process, improving material grading accuracy and production efficiency. In summary, the material grading deviation matrix constructed in step S3 provides data support and analysis tools for the airflow control and performance optimization of the non-powered powder classifier, reducing errors caused by human operation and improving the accuracy and efficiency of material grading.

[0089] S4. Input the material classification deviation matrix into the pre-constructed classifier classification accuracy evaluation model to obtain the material classification accuracy index corresponding to the non-powered classifier.

[0090] Step S4 is to evaluate the classification accuracy of the non-powered air classifier. This is done by inputting the material classification deviation matrix into a pre-built classification accuracy evaluation model to calculate an accuracy index that comprehensively reflects the classification effect. The specific steps are as follows:

[0091] The process of constructing the classification accuracy evaluation model for air classifiers includes:

[0092] Data preparation: Historical data is collected, and the material classification deviation matrix corresponding to the historical data is used as the input feature of the classifier classification accuracy evaluation model. The material classification deviation matrix represents the deviation between the actual classification effect and the target classification effect, and is information for evaluating classification accuracy. The material classification accuracy index corresponding to the material classification deviation matrix is ​​defined as the output label of the model. The material classification accuracy index is a quantitative indicator that reflects the overall quality of the classification effect.

[0093] Model selection: Based on the problem of quantifying the grading effect, an appropriate machine learning algorithm is selected to construct the grading accuracy evaluation model of the air classifier; including linear regression, support vector machine, random forest and neural network, etc.; historical data, namely the material grading deviation matrix and the corresponding grading accuracy index, are used as the training dataset to train the air classifier grading accuracy evaluation model to learn the relationship between input features and output labels.

[0094] Model training: Train the selected algorithm using the training dataset, and adjust the parameters of the classifier grading accuracy evaluation model until the classifier grading accuracy evaluation model can fit the training data well; use independent validation and test sets to evaluate the generalization ability of the classifier grading accuracy evaluation model to ensure that the classifier grading accuracy evaluation model can give accurate grading accuracy index predictions even on unseen data.

[0095] Model Application: The air classifier classification accuracy evaluation model receives a material classification deviation matrix as input. Each row in the matrix represents a set of deviation feature values ​​at a collection time point, and each column represents the change of deviation feature values ​​of different particle size levels over time. The air classifier classification accuracy evaluation model outputs a material classification accuracy index, which quantifies the classification effect of the non-powered air classifier under specific operating conditions.

[0096] Dynamic adjustment: During the production process, the classifier's classification accuracy evaluation model receives the latest material classification deviation matrix in real time and calculates the classification accuracy index, providing a basis for the dynamic adjustment of air volume control parameters. The classification accuracy index is used as a feedback signal and combined with the preset air volume control parameter database to find the optimal air volume control parameters, thereby optimizing the classification effect.

[0097] Performance optimization and updates: With the accumulation of more data, the air classifier classification accuracy evaluation model is updated regularly to adapt to changes in material properties, environmental conditions and equipment status, and to continuously optimize the predictive ability of the classification accuracy index; the predictive accuracy of the air classifier classification accuracy evaluation model is evaluated regularly, and the model is tuned or retrained when necessary to ensure its effectiveness and accuracy.

[0098] After the classification accuracy evaluation model of the air classifier is constructed and verified, the material classification accuracy index of the non-powered air classifier is calculated. The steps are as follows:

[0099] Prepare a sample of materials for grading: Select representative materials to ensure that the selected materials can represent the common material types and particle size distributions in actual production; Prepare standard grading conditions by defining a set of standard grading conditions, such as wind speed and feed rate, to facilitate grading under consistent conditions.

[0100] Conduct a grading experiment: Run the classifier under selected conditions to grade the prepared material sample; collect the graded material from each outlet, and the material is divided into coarse, fine and intermediate grades;

[0101] Particle size distribution determination: Use particle size analysis equipment, such as laser diffractometer and sieve tester, to determine the particle size distribution of each part of the material after classification;

[0102] Data processing: Record the raw data, including the particle size distribution of the material before grading and the particle size distribution of each part after grading; standardize the data by expressing all data using the same particle size level and unit to facilitate subsequent comparisons;

[0103] Calculate deviation: Compare the graded granularity distribution with the expected target granularity distribution to calculate the deviation; quantify the deviation: use appropriate methods, such as root mean square error (RMSE) and mean absolute error (MAE), to quantify the degree of deviation at each granularity level.

[0104] Application of the accuracy evaluation model: Deviation data is input into the air classifier's classification accuracy evaluation model. The model outputs an index representing the material classification accuracy based on the input deviation data. The material classification accuracy index ranges from 0 to 1; a higher value indicates a better classification effect.

[0105] Analysis Results: The obtained material classification accuracy index was analyzed to determine the classification performance of the air classifier under given conditions. The material classification accuracy index was lower than expected, indicating that potential problems in the classification process need to be identified, and it is necessary to consider whether the classification conditions need to be adjusted or the air classifier design needs to be optimized.

[0106] Adjustment and verification: Based on the analysis results, make necessary adjustments to the operating parameters of the air classifier, such as changing the air volume and adjusting the rotation speed; conduct the classification experiment again to confirm the effect of the adjustment until a satisfactory material classification accuracy index is achieved;

[0107] By collecting material sample data before and after grading, a grading accuracy evaluation model for the air classifier is constructed based on this data. This reduces reliance on manual experience and improves the objectivity and accuracy of grading accuracy evaluation. The model calculates the impact of different operating conditions on the grading effect, facilitating more refined management of the air classifier and improving grading performance. The model provides real-time evaluation of grading results, quickly identifies deviations from ideal conditions, and adjusts operating parameters promptly, enhancing the stability and efficiency of the grading process. Training and validation of the model allow for continuous optimization, improving grading accuracy, adapting to changes in the production environment, and increasing air classifier efficiency. Analyzing trends in grading performance helps predict potential equipment failures or operational errors, allowing for proactive measures to reduce the risk of production interruptions. The model adapts to different material types and production lines by adjusting parameters or retraining, demonstrating strong flexibility and adaptability. Precise control of the grading process reduces resource waste and human error caused by poor grading results, improving the quality of the final product and the grading accuracy of the air classifier.

[0108] S5. Based on the material classification accuracy index, the optimal air volume control parameters corresponding to the non-powered air classifier are mapped from the preset air volume control parameter database.

[0109] Step S5 uses the material classification accuracy index as a feedback signal to select the most suitable airflow setting from the preset airflow control parameter database to optimize the classification effect of the non-powered air classifier; the specific airflow control parameters include:

[0110] Air inlet opening: The size of the air inlet opening determines the airflow into the air classifier. By adjusting the air inlet opening, the air volume can be indirectly controlled, thereby affecting the material classification.

[0111] Air outlet opening: The opening of the air outlet also affects the air volume distribution and air pressure. Appropriate adjustment of the air outlet opening helps to optimize the airflow distribution in the air classifier and improve the classification effect.

[0112] Guide vane angle: The air classifier is equipped with adjustable guide vanes. Adjusting their angle will affect the air volume and direction, thus affecting the classification of materials.

[0113] Air pressure: Maintaining appropriate air pressure inside the air classifier is the basis for ensuring the classification effect. Too low or too high air pressure will affect the suspension state of the material, and thus affect the classification accuracy.

[0114] Airflow compensation factors: The difference in temperature and humidity inside and outside the classifier hopper will affect the actual effect of airflow. By adjusting the difference in temperature and humidity inside and outside, the airflow can be adjusted to maintain the stability of the classification effect.

[0115] First, construct a database of airflow control parameters, including:

[0116] List all variables that affect the classification effect, including inlet opening, outlet opening, guide vane angle, air pressure, and temperature and humidity inside and outside the classifier hopper; design an experimental matrix that covers all variables that affect the classification effect to ensure that the experiment covers a wide range of operating conditions.

[0117] Perform experiments according to the experimental plan, ensuring that only one variable is changed in each experiment while other variables remain unchanged to ensure data comparability; record detailed parameter settings and results for each experiment, including material classification accuracy index and other relevant performance indicators;

[0118] Inspect and clean the data, removing outliers and erroneous data points; convert the data to a uniform format, such as converting all air volume data to the same unit; store the cleaned data in a structured database to ensure data integrity and accessibility;

[0119] Design the database structure, including tables, fields, and data types, to ensure effective storage and retrieval of experimental data; define the relationships between different tables, such as the relationship between airflow and grading effect; use the database management module to create an airflow control parameter database and import the preprocessed data;

[0120] Analyze the data using statistical analysis or data mining techniques to identify factors that affect the grading effect; based on the data analysis results, establish a predictive model to predict the grading effect under different airflow settings, such as the material grading accuracy index.

[0121] Create indexes for key fields in the airflow control parameter database to improve query speed; optimize query statements to ensure that the optimal airflow control parameters can be retrieved quickly; if there is no directly matching record in the airflow control parameter database, use a pre-trained grading accuracy evaluation model to predict the grading effect under different airflow control parameters and select the generated airflow control parameters; after selecting the airflow control parameters, verify the effect through small-scale experiments or simulations to ensure that the expected grading accuracy is achieved in actual production;

[0122] Develop APIs or interfaces that allow direct access to the airflow control parameter database and adjust the optimal airflow control parameters in real time based on the graded accuracy index.

[0123] Regularly collect new data and update the air volume control parameter database to ensure it reflects the latest operating conditions and classification effects; continuously monitor the performance of the air volume control parameter database, and regularly optimize and maintain it to ensure stability and efficiency;

[0124] By collecting extensive experimental data and constructing an airflow control parameter database, decisions are made based on actual classification effect data, improving the accuracy and reliability of decision-making. The construction and maintenance of the airflow control parameter database allows for real-time or periodic updates of optimal airflow control parameters, ensuring that the non-powered classifier can adapt to changes in material characteristics and environmental conditions, achieving dynamic optimization. Precise airflow control improves classification efficiency, reduces energy consumption, and simultaneously reduces material waste and reprocessing costs caused by inaccurate classification. Regular updates to the airflow control parameter database and predictive models continuously optimize classification effects, improving product quality and production efficiency, and adapting to changes in production demands. Maintaining the airflow control parameter database provides historical records of operating parameters and classification effects, facilitating the tracking and analysis of root causes of problems and improving the transparency of the production process. Adjusting predictive model parameters and experimental matrices allows for flexible responses to changes in the production environment. Real-time monitoring of the classification accuracy index and comparison with the optimal parameters in the database enables closed-loop control, ensuring that the classification effect is always at its best. By constructing and utilizing the airflow control parameter database, precise control and continuous optimization of the non-powered classifier's classification effect are achieved, improving production efficiency and product quality while reducing operating costs.

[0125] S6. Control the air volume of the non-powered air classifier according to the optimal air volume control parameters;

[0126] Step S6 is the execution stage of the airflow control method for the non-powered air classifier. Based on the determined optimal airflow control parameters, the airflow of the non-powered air classifier is adjusted to optimize the classification effect. The detailed steps are as follows:

[0127] Ensure communication with the airflow control parameter database and read the optimal airflow control parameters; based on the received optimal parameters, generate specific airflow adjustment instructions;

[0128] The air volume of the non-powered air classifier is adjusted by the actuator to achieve the optimal air volume control parameters; during the production process, the air volume is dynamically adjusted according to the real-time monitored material classification accuracy index to adapt to changes in material characteristics and environmental conditions.

[0129] Deploy a real-time monitoring module to continuously track the grading effect after airflow adjustment, ensuring that the grading accuracy index reaches or approaches the optimal value; the monitoring module should have anomaly detection function, be able to identify situations where the grading effect after airflow adjustment does not meet expectations, and issue alarms in a timely manner;

[0130] Real-time monitoring of the grading effect ensures continuous optimization of airflow adjustment; through continuous adjustment and monitoring, it learns and adapts to different operating conditions, improving the stability of the grading effect.

[0131] Set a safety threshold; when the air volume adjustment exceeds the safe range, trigger an emergency shutdown procedure to avoid equipment damage or safety accidents; allow operators to manually intervene in the air volume adjustment to deal with emergencies or special needs.

[0132] Record the time, parameters, and results of each airflow adjustment for later analysis and optimization; regularly generate performance reports to summarize the effectiveness of the airflow control methods and provide a basis for improvement.

[0133] Regularly check the actuator's functionality to ensure it is operating normally;

[0134] By reading optimal airflow control parameters from the database and generating airflow adjustment commands, the inaccuracy and inefficiency of manual adjustments are reduced. Airflow adjustment via actuators ensures rapid and accurate adjustments, reducing response time and human error. Dynamic airflow adjustment allows the classifier to adapt to changes in material characteristics and environmental conditions in real time, maintaining stable classification effects and high classification accuracy. The deployment of a real-time monitoring module ensures continuous tracking of classification effects, facilitating immediate problem detection and corrective measures, reducing product quality degradation due to classification deviations. Anomaly detection enhances safety, reacting quickly when classification effects deviate from targets, reducing potential losses. A closed-loop control mechanism achieves continuous optimization through feedback loops, ensuring the accuracy of airflow adjustment and the stability of classification effects. Setting safety thresholds and emergency shutdown procedures prevents equipment overload or malfunction, protecting the equipment and operator safety. The generation of operation logs and performance reports provides detailed analytical data, helping technicians understand and optimize control strategies, promoting continuous improvement. Intelligent and automated methods improve the airflow control level of the non-powered classifier; precise airflow control enhances classification accuracy and production efficiency, while also improving overall production efficiency and reliability.

[0135] Example 2: Figure 3 As shown, the air volume control system for the non-powered air classifier of the present invention specifically includes the following modules;

[0136] The sample acquisition module is used to collect various material classification samples during the classification process of the non-powered air classifier;

[0137] The feature extraction module is used to analyze each material grading sample through preset material grading target features, calculate and extract the difference between the actual mass ratio of the target particle size material and the target set value, and generate a material grading deviation vector;

[0138] The deviation matrix generation module is used to collect and organize material classification deviation vectors at multiple time points to form a time series material classification deviation matrix. The columns of the material classification deviation matrix represent the deviation feature values ​​at different time points, and the rows represent the deviation feature values ​​of different target granularities at the same time point.

[0139] The classification accuracy evaluation module receives the deviation matrix as input and uses a pre-built classification accuracy evaluation model of the air classifier to calculate the material classification accuracy index of the non-powered air classifier.

[0140] The parameter mapping module is used to search and map the optimal air volume control parameters that best match the non-powered air classifier based on the calculated material classification accuracy index in the preset air volume control parameter database.

[0141] The execution module receives the optimal airflow control parameters from the airflow control parameter mapping module, dynamically adjusts the airflow of the non-powered air classifier, and optimizes the grading and classification effect.

[0142] By automatically collecting samples, analyzing data, evaluating accuracy, and adjusting parameters, the system automates airflow control, avoiding the uncertainty of manual adjustments based on experience and improving control precision and stability. The construction of material classification deviation vectors and matrices enables real-time monitoring of classification effects, timely detection of deviations from target settings, rapid response and adjustment, ensuring consistent classification accuracy. The classification accuracy evaluation model analyzes data and calculates a material classification accuracy index, ensuring that airflow adjustments are based on objective data rather than subjective judgment, improving the scientific rigor and reliability of decision-making. Dynamic airflow adjustment more effectively achieves the desired classification effect, improving material classification accuracy and enhancing... This system improves the quality of cement and other products and the overall efficiency of the production line; automated and precise airflow control reduces material rework and loss caused by inaccurate grading, lowers production costs, and saves energy; it automatically adjusts airflow control parameters according to different material properties and environmental conditions, improving the equipment's adaptability and operational flexibility under varying working conditions; as more data is collected, the grading accuracy evaluation model is continuously optimized through machine learning technology, improving its predictive and evaluation capabilities and achieving continuous system improvement; through automated and real-time control, it reduces human intervention and downtime for adjustments, improving the continuity, stability, product quality, and production efficiency of the production line.

[0143] The various variations and specific embodiments of the non-powered air classifier airflow control method in the aforementioned Embodiment 1 are also applicable to the non-powered air classifier airflow control system of this embodiment. Through the foregoing detailed description of the non-powered air classifier airflow control method, those skilled in the art can clearly understand the implementation method of the non-powered air classifier airflow control system in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0144] 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.

[0145] 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 controlling the air volume of a non-powered air classifier, characterized in that, The method includes: Several material classification samples were collected from the non-powered air classifier. Based on the preset material classification target features, target deviation features are extracted for each material classification sample to obtain the material classification deviation vector corresponding to the collection time. The material classification deviation vector consists of several deviation feature values, each of which corresponds to a material classification sample. The material classification target features include the target particle size and the target particle size mass ratio corresponding to each material classification sample. The deviation feature value refers to the difference between the mass ratio of the target particle size material and the target particle size mass ratio in the material classification sample. Obtain material classification deviation vectors from multiple acquisition times and arrange them in chronological order to obtain a material classification deviation matrix. In this matrix, deviation feature values ​​corresponding to the same target granularity from different acquisition times are located in the same column, and deviation feature values ​​corresponding to different target granularities acquired at the same time are located in the same row. The material classification deviation matrix is ​​as follows: ； in, It is an element in the material classification deviation matrix, representing the first element in the material classification deviation matrix. The first collection time The deviation eigenvalue at the granularity level; the first The row contains deviation eigenvalues ​​at all granularity levels, corresponding to the first row. The first data collection time point; the first The column contains the number of data collection points at all time points. Deviation eigenvalues ​​at each granularity level; The material classification deviation matrix is ​​input into a pre-constructed air classifier classification accuracy evaluation model to obtain the material classification accuracy index corresponding to the non-powered air classifier. The construction method of the air classifier classification accuracy evaluation model includes: collecting information on the material before and after classification by the non-powered air classifier under different airflow control parameter settings as the training basis for the air classifier classification accuracy evaluation model; determining the target particle size range and the particle size distribution curve of the target mass ratio of each particle size segment; extracting features from the classified material, including the percentage of material in different particle size segments, the recovery ratio of coarse and fine powder, and the recovery rate of fine powder minus the proportion of fine powder in coarse powder; calculating the deviation between the actual particle size distribution and the target particle size distribution; selecting a machine learning algorithm to construct the air classifier classification accuracy evaluation model; using the material classification deviation matrix and the corresponding classification accuracy index in historical data as the training dataset to train the air classifier classification accuracy evaluation model; adjusting the parameters of the air classifier classification accuracy evaluation model; and using independent validation and test sets to evaluate the generalization ability of the air classifier classification accuracy evaluation model. Based on the material classification accuracy index, the optimal air volume control parameters corresponding to the non-powered air classifier are mapped from the preset air volume control parameter database. The air volume of the non-powered air classifier is controlled according to the optimal air volume control parameters.

2. The airflow control method for a non-powered air classifier as described in claim 1, characterized in that, Methods for collecting classification samples of several materials obtained from classification using a non-powered air classifier include: Determine and sample the locations of the non-powered air classifier; the sampling locations include the feed inlets and outlets of different levels of the air classifier. The sampling frequency and cycle are determined based on production needs and equipment operating conditions; Select sampling tools and equipment for sampling; The sampling method is selected based on the location of the sampling points and the properties of the material; sampling methods include continuous flow sampling, manual-assisted sampling, and mechanical-assisted sampling. After sampling, each collected material graded sample is identified and recorded, including sampling time, sampling point, and sample number information.

3. The airflow control method for a non-powered air classifier as described in claim 1, characterized in that, The method for constructing the air volume control parameter database includes: List the airflow control parameters that affect the grading effect; design an experimental matrix, which includes the airflow control parameters that affect the grading effect. Perform the experiment by changing only one airflow control parameter while keeping the other airflow control parameters unchanged. Repeat the experiment multiple times according to the number of airflow control parameters and record the parameter settings and results of each experiment. Check and clean the experimental data, remove outliers and erroneous data points, convert the cleaned experimental data into a uniform format, and store the converted experimental data in the database; Design a database for storing and querying experimental data. The database structure includes tables, fields, and data types. Use the database management module to create a database of air volume control parameters and import the pre-processed data.

4. The airflow control method for a non-powered air classifier as described in claim 3, characterized in that, The air volume control parameters include the air inlet opening, air outlet opening, guide vane angle, air pressure, and the temperature and humidity inside and outside the classifier hopper.

5. A non-powered airflow control system for a classifier, characterized in that, The system is applied to the airflow control method for a non-powered air classifier as described in claim 1, and the system includes: The sample acquisition module is used to collect various material classification samples during the classification process of the non-powered air classifier; The feature extraction module is used to analyze each material grading sample through preset material grading target features, calculate and extract the difference between the actual mass ratio of the target particle size material and the target set value, and generate a material grading deviation vector; The deviation matrix generation module is used to collect and organize material classification deviation vectors at multiple time points to form a time series material classification deviation matrix. The columns of the material classification deviation matrix represent the deviation feature values ​​at different time points, and the rows represent the deviation feature values ​​of different target granularities at the same time point. The classification accuracy evaluation module receives the deviation matrix as input and uses a pre-built classification accuracy evaluation model of the air classifier to calculate the material classification accuracy index of the non-powered air classifier. The parameter mapping module is used to search and map the optimal air volume control parameters that best match the non-powered air classifier based on the calculated material classification accuracy index in the preset air volume control parameter database. The execution module receives the optimal airflow control parameters from the airflow control parameter mapping module, dynamically adjusts the airflow of the non-powered air classifier, and optimizes the grading and classification effect.

6. An electronic device for controlling the airflow of a non-powered air classifier, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that, When the computer program is executed by the processor, it implements the steps of the method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-4.

Images