Wind power screening method for dry-method sand making building based on AI identification

Abstract

The invention relates to a dry-method sand making building wind power screening method based on AI recognition, and belongs to the technical field of gravel aggregate production, and the dry-method sand making building wind power screening method based on AI recognition comprises the steps that S1, sand material information in the wind power screening process is obtained; s2, identifying and analyzing the sand material information by using an AI identification module; and S3, sand making parameters of the dry-method sand making building are regulated and controlled based on the analysis result. Key data such as the particle state, the dust concentration and the equipment operation state in the screening process are collected and analyzed in real time through the AI recognition module, wind power screening parameters and related equipment operation states are dynamically regulated and controlled in combination with preset standards and historical data, intelligent regulation and control of wind power screening are achieved, the screening quality is improved, and screening energy consumption is reduced.

Description

AI-based wind screening method for dry sand making towers Technical Field

[0001] This application relates to the field of sand and gravel aggregate production, and in particular to a wind-powered screening method for dry sand making towers based on AI recognition. Background Technology

[0002] Dry sand making towers, as an energy-saving and environmentally friendly sand and gravel aggregate production equipment, rely on the core technologies of crushing and air screening. With the advantage of not requiring the addition of water throughout the process, this equipment effectively avoids many drawbacks of traditional wet sand making processes, such as high water consumption, large wastewater discharge, excessive moisture content in sand, and inability to produce in winter due to freezing. It has become one of the mainstream equipment in the sand and gravel aggregate production industry.

[0003] Specifically, the production process of a dry sand making plant is roughly as follows: raw stone (such as granite, basalt, limestone, etc.) is fed into a high-pressure crusher by a conveying device. Through the squeezing and impact of the hammers, liners and other components inside the high-pressure crusher, the raw stone is crushed into stones of a certain particle size range. The crushed stones are then conveyed into the air screening area by a conveying device. The airflow generated by the blower blows away the dust, fine particles that do not meet the size requirements and coarse particles mixed in with the stones, leaving only the sand and gravel aggregate that meets the preset standards. Finally, the standard sand and gravel aggregate is transported to the qualified aggregate collection bin to complete the production of sand and gravel aggregate.

[0004] It is evident that the screening effect of the wind screening process directly determines the quality, production efficiency, and environmental performance of the finished sand and gravel aggregates, making it a key and core step in the dry sand production process.

[0005] However, the air screening process in existing dry sand making plants still faces numerous technical bottlenecks, severely restricting further improvements in production efficiency, product quality, and environmental performance. Specific pain points include: First, the air screening parameters are fixed, making dynamic adjustment impossible. In existing dry sand making plants, the core screening parameters such as wind speed and air volume are mostly manually preset fixed values, or only a few fixed settings are available for manual switching. This makes dynamic adjustment impossible based on raw material characteristics (such as the hardness, composition, and initial particle size distribution of the raw stone), the crushing effect of the high-pressure crusher (the actual particle size distribution and uniformity of the crushed stone), and real-time production needs. This fixed-parameter screening method results in low screening accuracy. When the wind speed and air volume are set too high, some qualified particles that meet the requirements will be mistakenly blown away; when the wind speed and air volume are set too low, dust and fine particles that do not meet the size requirements mixed in the stone cannot be completely separated, leading to excessive powder content and unqualified particle size distribution in the finished sand and gravel aggregate.

[0006] Secondly, the screening process is highly unpredictable, resulting in significant energy consumption and environmental pressure. In the existing wind-powered screening process, the separation of dust, qualified particles, and unqualified particles relies entirely on the action of wind power. There is a lack of a precise identification and directional separation mechanism for the three types of substances. This requires the fans to maintain high power operation for a long time in order to achieve the best possible separation effect, resulting in high energy consumption in wind-powered screening and increasing the production and operating costs of enterprises.

[0007] Third, the identification of abnormal situations is lagging, relying on manual inspections, resulting in low management efficiency. Existing screening methods cannot identify various abnormal situations during the screening process in real time, including particle agglomeration (clumping of crushed stones due to humidity, viscosity, etc., causing blockage of the screening channel), screening channel blockage (accumulation of agglomerated or coarse particles, causing poor material conveying), and particle size abnormalities caused by wear of the high-pressure crusher (wear of the crusher's hammers and liners reduces crushing efficiency, causing the particle size distribution of the crushed stones to deviate from the preset range). The investigation of these abnormal situations relies entirely on manual inspections, which not only requires a significant investment of manpower but also suffers from untimely investigations and inaccurate judgments. Failure to address abnormalities promptly can lead to equipment failures and production interruptions, as well as fluctuations in product quality, affecting the stability of the finished sand and gravel aggregates.

[0008] Fourth, the adjustment of screening parameters lacks data support and has insufficient adaptive optimization capabilities. Existing screening systems rely heavily on operator experience for parameter adjustments, lacking scientific support based on real-time particle state, dust concentration, and other data, thus failing to achieve adaptive optimization of screening parameters. When raw material characteristics change, high-pressure crushers experience wear, or production demands are adjusted, operators cannot quickly and accurately adjust the air-powered screening parameters, leading to unstable screening results and difficulty in adapting to the production needs of different raw material characteristics and different specifications of sand and gravel aggregates. Summary of the Invention

[0009] To improve the intelligence level of wind screening in dry sand making plants, this application provides a wind screening method for dry sand making plants based on AI recognition.

[0010] In one aspect of this disclosure, an AI-based method for wind-driven screening of dry sand making plants is proposed, comprising: S1, acquiring sand information during the wind-driven screening process; S2, using an AI recognition module to identify and analyze the sand information; and S3, adjusting the sand making parameters of the dry sand making plant based on the analysis results.

[0011] Preferably, the sand information includes sand images, particle size, and dust concentration data within the wind-powered screening channel.

[0012] Preferably, in step S2, the AI ​​recognition module is trained before it performs recognition and analysis on the sand information. The training of the AI ​​recognition module includes: S21, acquiring stone samples with different raw material characteristics, different crushing particle sizes, and different dust contents during the dry sand making process, as well as individual and mixed samples of qualified sand and gravel aggregates, unqualified fine particles, unqualified coarse particles, and dust; S22, collecting image data, particle size data, and dust concentration data of the samples; S23, preprocessing and labeling the sample data, and inputting the preprocessed sample data into a deep convolutional neural network to train an AI recognition module capable of particle type recognition, particle size detection, and dust concentration recognition.

[0013] Preferably, the sample data in S23 also includes crushing sample data under different wear levels of the sand making equipment and screening sample data under different wind parameters.

[0014] Preferably, S3, adjusting the sand-making parameters of the dry sand-making plant based on the analysis results includes: determining whether it is necessary to adjust the sand-making parameters of the dry sand-making plant based on the analysis results; if it is determined that adjustment is necessary, then adjust the sand-making parameters of the dry sand-making plant; if it is determined that adjustment is not necessary, then keep the current sand-making parameters of the dry sand-making plant unchanged.

[0015] Preferably, when it is determined that adjustment is needed, the sand making parameters of the dry sand making tower are adjusted, and the wind screening method further includes: obtaining sand information during the wind screening process after the sand making parameters are adjusted; repeating S2 and S3.

[0016] Preferably, S1 further includes acquiring the equipment status and operating data information of the dry sand making plant; S2 further includes using an AI recognition module to identify and analyze the equipment status and operating data information of the dry sand making plant.

[0017] Preferably, the sand making parameters include the wind speed and air volume of the wind screening, and the crushing parameters of the sand making equipment.

[0018] Beneficial technical effects: By using an AI recognition module to collect and analyze key data such as particle state, dust concentration, and equipment operating status in real time during the screening process, and combining preset standards and historical data, the system can dynamically adjust wind screening parameters and related equipment operating status to achieve intelligent control of wind screening, improve screening quality, and reduce screening energy consumption. Detailed Implementation

[0020] In one aspect of this disclosure, a method for wind-driven screening of dry sand making towers based on AI recognition is proposed, comprising: S1, acquiring sand information during the wind-driven screening process; S2, using an AI recognition module to identify and analyze the sand information; and S3, adjusting the sand making parameters of the dry sand making tower based on the analysis results.

[0021] Specifically, the sand information in S1 includes sand images, particle size, and dust concentration data within the wind-powered screening channel.

[0022] Images of the sand can be captured by cameras installed at the feed end, middle, and discharge end of the air-powered screening channel.

[0023] Particle size can be detected and obtained by particle size sensors installed at the feed end, middle and discharge end of the air screening channel. Particle size sensors include two main categories: laser scattering particle size sensors and machine vision particle size sensors.

[0024] Dust concentration is detected and obtained by dust concentration sensors, which are installed at the feed end, middle and discharge end of the wind-powered screening channel.

[0025] Furthermore, S1 also includes acquiring the equipment status and operating data information of the dry sand making plant, so that the AI ​​recognition module can promptly identify abnormal equipment status of the dry sand making plant.

[0026] Specifically, the equipment status of the dry sand making plant includes the operating status of the crushing equipment, the air generation equipment, the screening channel, and the dust removal equipment. Operational data includes whether there is particle clumping or blockage in the screening channel, whether the dust concentration exceeds the standard, the wear condition of the crushing equipment, and the dust removal efficiency of the dust removal equipment.

[0027] It should be noted that after starting the dry sand making plant, the raw stone is fed into the crushing equipment (such as a high-pressure crusher) through a conveyor device. The raw stone is crushed according to the preset initial operating parameters to the preset initial particle size range (0~5mm). The crushed stone is then conveyed to the feed end of the air-powered screening channel by a belt conveyor to ensure that the material enters the screening channel evenly and continuously, avoiding material accumulation that could lead to poor screening results.

[0028] Simultaneously, the wind power generator (variable frequency fan) and pulse bag filter are activated, and a stable airflow is introduced into the wind screening channel according to the preset initial values ​​of wind power parameters and dust removal parameters to carry out preliminary wind screening. The purpose of preliminary screening is to quickly separate most of the dust and fine unqualified particles mixed in the crushed stones, reducing the load on subsequent precision screening. The dust separated in the preliminary screening is collected by the pulse bag filter and sent to the dust collection bin. The material after preliminary screening continues to move in the wind screening channel and enters the subsequent precision screening stage. In this embodiment, S2 uses an AI recognition module to identify and analyze the sand information, thereby identifying the distribution and proportion of qualified particles, unqualified fine particles, unqualified coarse particles, and dust, and determining whether the current particle size distribution meets the preset standard.

[0029] Furthermore, it further determines whether the current dust concentration exceeds a preset threshold based on dust concentration data. It also assesses the operational status of crushing equipment, wind power generation equipment, dust removal equipment, etc., based on equipment operation data, identifying any abnormalities such as equipment wear or malfunctions. Simultaneously, it identifies any abnormalities in the screening process, such as particle agglomeration or blockage of screening channels.

[0030] It should be noted that before the AI ​​recognition module in S2 can identify and analyze the sand information, it needs to be trained so that it can identify abnormalities based on the sand information acquired during the wind screening process.

[0031] Specifically, the training of the AI ​​recognition module includes: S21, acquiring stone samples with different raw material characteristics, different crushing particle sizes, and different dust contents during the dry sand making process, as well as individual and mixed samples of qualified sand and gravel aggregates, unqualified fine particles, unqualified coarse particles, and dust; specifically, by carrying out comprehensive sample collection work, sufficient data support is provided for the training of the AI ​​recognition module. The collection scope covers various scenarios that may be encountered in the dry sand making process, specifically including: raw stone samples with different raw material characteristics (such as granite, basalt, limestone, etc. with different hardness and composition); crushed stone samples with different crushing effects (such as different particle size distribution and different particle uniformity); stone samples with different dust content; and individual and mixed samples of qualified sand and gravel aggregate (compliant with GB / T 14684-2022 "Construction Sand" standard, particle size range 0.15~4.75mm, dust content 3%~7%), unqualified fine particles (particle size <0.15mm), unqualified coarse particles (particle size >4.75mm), and dust.

[0032] S22. Collect image data, particle size data, and dust concentration data of the samples; when the crushing equipment is a high-pressure crusher and the wind power generating equipment is a fan, the collected sample data includes high-definition image data, particle size data, and dust concentration data of various samples; at the same time, collect crushed sample data under different wear levels of the high-pressure crusher and screened sample data under different wind parameters for subsequent module optimization.

[0033] S23. Preprocess and label the sample data, and input the preprocessed sample data into a deep convolutional neural network to train an AI recognition module with particle type recognition, particle size detection, and dust concentration recognition.

[0034] Furthermore, the sample data in S23 also includes crushing sample data under different wear levels of the sand making equipment and screening sample data under different wind parameters.

[0035] Correspondingly, the AI ​​recognition module can also identify and analyze the equipment status and operating data of the dry sand making plant, so that the AI ​​recognition module can quickly identify abnormal equipment conditions in the dry sand making plant.

[0036] Specifically, the preprocessing of the collected sample data includes: First, denoising (removing light and dust interference during the shooting process), enhancement (improving image contrast and clarity), cropping, and labeling of the image data. The labeling content clearly includes particle type (qualified sand and gravel aggregate, unqualified fine particles, unqualified coarse particles, dust), particle size range, dust concentration level, and whether particle agglomeration exists. Outlier removal and standardization are performed on particle size data and dust concentration data to ensure the accuracy and consistency of the data.

[0037] Secondly, the preprocessed sample data was divided into training, validation, and test sets in a 7:2:1 ratio, and then input into a deep convolutional neural network (CNN) to train, validate, and test the AI ​​recognition module. During training, the network parameters of the model (such as kernel size, learning rate, and number of iterations) were continuously adjusted to optimize the model's recognition accuracy and response speed, ultimately obtaining an AI recognition model that meets production requirements.

[0038] The model must possess four core functions: First, accurate particle type identification, capable of quickly distinguishing between qualified sand and gravel aggregates, unqualified fine particles, unqualified coarse particles, and dust, with an accuracy rate ≥98%; second, particle size detection, capable of real-time detection of the actual particle size distribution, with a detection error ≤0.02mm; third, dust concentration identification, capable of accurately determining the current dust concentration level based on image data and sensor data, with an identification error ≤0.2mg / m³; and fourth, abnormal situation identification, capable of real-time identification of particle agglomeration, screening channel blockage, excessive dust concentration, and particle size abnormalities caused by wear of the high-pressure crusher, with an identification response time ≤50ms.

[0039] In this embodiment, after S2 activates the AI ​​recognition module, it uses equipment status detection devices such as high-definition cameras, particle size sensors, and dust concentration sensors to collect various key data in the screening process in a comprehensive and real-time manner, ensuring the comprehensiveness, accuracy, and real-time nature of data collection.

[0040] Among them, high-definition cameras capture the movement state and particle distribution of materials in real time from the feed end, middle and discharge end of the wind-powered screening channel, capturing image information such as particle shape, size and agglomeration. Each frame of image is synchronously transmitted to the data storage and analysis module.

[0041] The particle size sensor detects the particle size distribution of the crushed stones at the feed end, the particle size distribution of the material in the middle, and the particle size distribution of the finished aggregate at the discharge end in real time, and transmits the detected particle size data to the data storage and analysis module in real time.

[0042] The dust concentration sensor detects the dust concentration inside the screening channel and at key locations in the workshop in real time, and transmits the concentration data to the data storage and analysis module in real time.

[0043] The equipment status monitoring device collects real-time equipment operation data such as the rotation speed and hammer wear of the high-pressure crusher, the power and rotation speed of the variable frequency fan, and the dust removal efficiency of the pulse bag filter, and transmits them to the data storage and analysis module.

[0044] The AI ​​recognition module performs real-time analysis and processing of the collected data. Based on image and particle size data, it identifies qualified particles, unqualified fine particles, unqualified coarse particles, and the distribution location and proportion of dust, determining whether the current particle size distribution meets preset standards. Based on dust concentration data, it determines whether the current dust concentration exceeds a preset threshold. Based on equipment operation data, it assesses the normal operating status of equipment such as the high-pressure crusher, variable frequency fan, and pulse bag filter, identifying any abnormalities such as equipment wear or malfunctions. Simultaneously, it identifies any abnormalities in the screening process, such as particle agglomeration or blockage of the screening channels.

[0045] In addition, the AI ​​recognition module generates recognition and analysis results and transmits the analysis results to the data storage and analysis module, which then adjusts the sand making parameters of the dry sand making plant based on the analysis results.

[0046] In this embodiment, S3, adjusting the sand-making parameters of the dry sand-making plant based on the analysis results includes: determining whether the sand-making parameters of the dry sand-making plant need to be adjusted based on the analysis results; if it is determined that adjustment is needed, then the sand-making parameters of the dry sand-making plant are adjusted; if it is determined that adjustment is not needed, then the current sand-making parameters of the dry sand-making plant are kept unchanged.

[0047] Furthermore, when it is determined that adjustment is needed, the air screening method further includes: obtaining sand information during the air screening process after the sand screening parameters are adjusted; repeating S2 and S3 until it is determined that no adjustment is needed, then keeping the current sand screening parameters of the dry sand making plant unchanged.

[0048] The sand making parameters include the wind speed and air volume of the wind screening, and the crushing parameters of the sand making equipment.

[0049] In this embodiment, the data storage and analysis module uses a built-in data processing algorithm to perform in-depth comparison and optimization of the analysis results, generating optimal control commands. The specific analysis process is as follows: First, the data storage and analysis module compares the real-time collected particle size distribution data with the preset sand and gravel particle size standard, calculates the particle size deviation value, and determines whether the current screening effect meets the standard. If the particle size deviation value is within the allowable range (≤±2%), the current wind force parameters are maintained unchanged. If the particle size deviation value exceeds the allowable range, such as a low percentage of qualified particles or too many unqualified particles remaining, it indicates that the current wind speed and air volume are insufficient, and the wind speed and air volume need to be appropriately increased. If too many qualified particles are mistakenly blown away, it indicates that the current wind speed and air volume are too high, and the wind speed and air volume need to be appropriately reduced.

[0050] Secondly, the real-time dust concentration data is compared with the preset dust concentration threshold. If the dust concentration is close to or exceeds the threshold, it indicates that the current dust removal effect is not good. The operating parameters of the pulse bag dust collector need to be adjusted through the dust removal coordination module, and the fan speed should be adjusted appropriately to reduce dust overflow.

[0051] Secondly, based on the equipment operation data, if it is found that the hammer of the high-pressure crusher is severely worn, resulting in abnormal particle size distribution after crushing, the operating parameters of the high-pressure crusher need to be adjusted (such as increasing the speed and reducing the feed rate). At the same time, the air screening parameters should be adjusted to adapt to the new crushing effect. If it is found that the variable frequency fan or pulse bag dust collector is operating abnormally, an abnormal signal should be issued in time to trigger the alarm module.

[0052] Finally, by combining historical production data, we analyzed the impact of current raw material characteristics and equipment operating status on the screening effect, and optimized the control parameters to ensure their rationality and suitability.

[0053] The data storage and analysis module sends the generated optimal control commands (including fan speed and air volume adjustment values, high pressure crusher operating parameter adjustment values, pulse bag dust collector operating parameter adjustment values, etc.) to the parameter control module of the dry sand making plant in real time, and the parameter control module adjusts the corresponding parameters in real time.

[0054] For example, by adjusting the frequency of the variable frequency fan, the wind speed and air volume can be precisely adjusted to ensure that a gradient airflow is formed in the screening channel (the bottom airflow is used to remove unqualified coarse particles, the middle airflow is used to retain qualified fine particles, and the top airflow is used to separate dust), thus achieving directional separation of particles and dust.

[0055] For example, adjusting the speed and feed rate of the high-pressure crusher can optimize the crushing effect and ensure that the particle size distribution after crushing meets the preset requirements.

[0056] In addition, the dust removal velocity and cleaning cycle of the pulse bag filter can be adjusted to enhance the dust removal effect, prevent dust from overflowing, and take into account the dust removal energy consumption.

[0057] After the parameters are adjusted, the AI ​​recognition module continues to collect sand information in real time during the wind screening process after the sand making parameters are adjusted. The AI ​​recognition model, data storage and analysis module continue to perform real-time analysis to verify the adjusted screening effect, equipment operating status and dust concentration, and determine whether the adjusted parameters have achieved the optimal effect.

[0058] If the adjusted screening effect meets the standard, the dust concentration meets the requirements, and the equipment operates normally, then maintain the current parameters.

[0059] If deviations still exist after adjustment (such as particle size deviation not being eliminated or dust concentration still exceeding the standard), repeat S3 and adjust the control parameters again until the screening effect, dust concentration, and equipment operating status all meet the preset requirements, forming a closed-loop optimization control of "collection-identification-analysis-control-verification" to ensure the stability and accuracy of the screening process.

[0060] This embodiment of the AI-based dry sand making tower wind-powered screening method introduces AI recognition technology to achieve precise perception and dynamic control of the screening process, breaking the limitations of existing fixed-parameter screening. The AI ​​recognition module collects data such as particle state, dust concentration, and equipment operating status in real time. Combined with the precise analysis of the AI ​​recognition module, the wind-powered screening parameters are adaptively adjusted to ensure that the screening effect always meets the preset standards, avoiding waste of qualified particles and residue of unqualified particles.

[0061] Simultaneously, a gradient airflow directional separation mechanism is constructed, combined with AI for precise identification, to achieve directional separation of dust, qualified particles, and unqualified particles, reducing screening energy consumption. By dynamically adjusting wind speed and air volume, a layered gradient airflow is formed within the screening channel, specifically separating different types of particles, avoiding the blind action of wind, effectively reducing fan operating energy consumption, and improving screening accuracy.

[0062] Furthermore, the AI-based dry sand making tower air screening method in this embodiment adaptively optimizes screening parameters according to changes in raw material characteristics and equipment operating status, adapting to different raw materials and different production needs, thereby improving the system's versatility and flexibility.

[0063] The following specific embodiment will be used to provide a detailed description of the AI-based dry sand making tower wind screening method of this application.

[0064] This embodiment is based on a dry sand making plant with a processing capacity of 15t / h. The specific equipment parameters are as follows: 1. Dry sand making plant body: High-pressure crusher model PF-1214, rated speed 1480r / min, feed particle size ≤300mm, discharge particle size 0~5mm; wind screening channel length 8m, diameter 1.2m, qualified aggregate collection bin volume 50m³, dust collection bin volume 20m³, variable frequency fan model 4-72-11, rated power 37kW, frequency conversion range 20~50Hz, air volume adjustment range 1500~3000m³ / h, wind speed adjustment range 5~12m / s, pulse bag dust collector model MC-96, dust removal efficiency ≥99.8%, dust cleaning cycle adjustable range 10~30min.

[0065] 2. An AI recognition module, three high-definition cameras (1080P resolution, 25 frames / s frame rate, 360° imaging without blind spots), two particle size sensors (0-5mm detection range, 0.01mm detection accuracy), three dust concentration sensors (0-50mg / m³ detection range, 0.1mg / m³ detection accuracy), and several equipment status sensors are installed on the high-pressure crusher, variable frequency fan, and pulse bag dust collector, respectively.

[0066] 3. Other modules: The parameter control module has a response delay of 0.3s; the data storage and analysis module has built-in data processing algorithms and historical production data; the alarm module has audible and visual alarms and mobile phone push functions; the dust removal coordination module is linked with the pulse bag dust collector and can adjust the dust removal parameters in real time.

[0067] The implementation steps are as follows: Sample collection and AI model training: Collect samples of granite and basalt, and collect stone samples with different crushing sizes (0~1mm, 1~3mm, 3~5mm) and different dust contents (3%, 5%, 7%, 10%, 12%), as well as individual and mixed samples of qualified sand and gravel aggregate (particle size 0.15~4.75mm, dust content 5%), unqualified fine particles (particle size <0.15mm), unqualified coarse particles (particle size >4.75mm), and dust; [The text abruptly ends here, likely due to an incomplete sentence or missing information.] The image data was denoised, enhanced, and labeled. Outlier removal and standardization were performed on particle size and dust concentration data. The preprocessed sample data was divided into training, validation, and test sets in a 7:2:1 ratio and input into a deep convolutional neural network (CNN) to train the AI ​​recognition model. The final model had a response time of 45ms, a particle type recognition accuracy of 98.3%, a particle size detection error of 0.015mm, and a dust concentration recognition error of 0.15mg / m³. The preset sand and gravel particle size standard conformed to GB / T 14684-2022 "Construction Sand" standard. The dust concentration threshold was ≤15mg / m³ in the workshop and ≤20mg / m³ in the screening channel. The initial wind parameters were wind speed of 7m / s and air volume of 1800m³ / h. The data was stored in the data storage and analysis module.

[0068] Material conveying and initial screening after crushing: The granite raw material is fed into the high-pressure crusher via a belt conveyor. The high-pressure crusher is started according to the preset initial parameters (speed 1480 r / min, feed speed 2 t / h) to crush the raw material to 0~5 mm. The crushed stone is then conveyed to the feed end of the air-powered screening channel via a belt conveyor. The variable frequency fan and pulse bag dust collector are started. According to the initial parameters of wind speed 7 m / s, air volume 1800 m³ / h, dust removal wind speed 1.2 m / s, and dust removal cycle 20 min, airflow is introduced into the screening channel for preliminary screening. Some dust and fine unqualified particles are initially separated. The dust is collected by the pulse bag dust collector and sent to the dust collection bin.

[0069] AI Real-Time Recognition and Data Acquisition: The AI ​​recognition module is activated, with three high-definition cameras capturing material images from the feed end, middle section, and discharge end. Particle size sensors detect the particle size distribution at these points in real time, while dust concentration sensors monitor dust concentration in the screening channel and workshop. Equipment status sensors collect real-time operating data from the high-pressure crusher, variable frequency fan, and pulse bag filter. The AI ​​recognition model analyzes the collected data in real time, yielding the following results: At the feed end, particle size distribution is 0-1mm (25%), 1-3mm (50%), and 3-5mm (25%), with a powder content of 12%; at the middle section, particle size distribution is 0-1mm (18%), 1-3mm (62%), and 3-5mm (20%), with a powder content of 8%; workshop dust concentration is 12mg / m³, and dust concentration in the screening channel is 18mg / m³; equipment operation is normal with no abnormalities. The collected data and analysis results are transmitted to the data storage and analysis module in real time.

[0070] Data processing and control parameter generation: The data storage and analysis module performs in-depth analysis of the collected data and analysis results, comparing them with preset standards: the powder content at the feed end is 12%, exceeding the preset threshold (≤10%), the powder content in the middle is 8%, close to the threshold, and the proportion of qualified particles does not meet the preset requirements (≥85%); combined with historical production data, the analysis concludes that the current wind speed and air volume are insufficient, and it is necessary to appropriately increase the wind speed and air volume, while adjusting the dust removal parameters; generating control commands: adjusting the variable frequency fan wind speed to 8.5m / s and the air volume to 2200m³ / h, adjusting the pulse bag dust collector dust removal wind speed to 1.5m / s and shortening the dust removal cycle to 15min, keeping the high-pressure crusher speed unchanged, and sending the control commands to the parameter control module.

[0071] Precise control and closed-loop optimization: After receiving control commands, the parameter control module adjusts the operating parameters of the variable frequency fan and pulse bag dust collector in real time. After the adjustment, the AI ​​recognition module collects and analyzes the data again, and obtains the following results: the particle size distribution in the middle is 0~1mm accounting for 10%, 1~3mm accounting for 75%, and 3~5mm accounting for 15%, with a dust content of 5.2%; the dust concentration in the workshop is 9mg / m³, and the dust concentration in the screening channel is 12mg / m³; the proportion of qualified particles is 97.3%, which meets the preset standard; the data storage and analysis module judges that the adjusted parameters have reached the optimal level, maintains the current parameter operation, and forms a closed-loop optimization.

[0072] Anomaly Handling and Alarms: During the screening process, if the AI ​​recognition module detects a small amount of particle clumping in the middle of the screening channel, the data storage and analysis module will immediately trigger the alarm module, issue an audible and visual alarm signal, and push the abnormal information "particle clumping, located in the middle of the screening channel, minor severity" to the operator's mobile terminal. After receiving the alarm, the operator will go to the site and use special tools to clean up the clumped particles. After cleaning, the operator will issue a recovery command through the workshop control platform, and the system will return to normal operation.

[0073] Finished Product Collection and Dust Recovery: Qualified sand and gravel aggregates, after screening, settle from the discharge end to the qualified aggregate collection bin, and are then conveyed by a belt conveyor to the finished product stockpile area, collecting 14.2 tons of qualified aggregates per hour. Unqualified coarse particles settle to the waste inlet and are fed into a high-pressure crusher for secondary crushing, processing 0.8 tons of waste per hour. Dust is collected by a pulse-jet bag filter, recovering 0.18 tons of dust per hour. The recovered dust is transported off-site for reuse as a concrete admixture. Implementation Results

[0074] After 72 hours of operation, the AI-based dry sand making tower wind screening method of this embodiment achieved the following results compared with the traditional fixed parameter screening process: 1. Screening accuracy and product quality: The powder content of the finished sand and gravel aggregate is stable at 5.0%~5.5%, the particle size distribution compliance rate is increased from 82% in the traditional process to 98.5%, and the rejection rate of unqualified particles is 99.2%, which fully complies with the GB / T 14684-2022 "Construction Sand" standard.

[0075] 2. Reduced energy consumption: The average energy consumption of the variable frequency fan decreased from 32kW / h to 25.2kW / h, a reduction of 21.2%; the average energy consumption of the pulse bag filter decreased from 8kW / h to 6.6kW / h, a reduction of 17.5%; and the energy consumption of the overall screening process decreased by 20.1%.

[0076] 3. Environmental performance: The dust concentration in the workshop is stable at 8~12mg / m³, and the dust concentration in the screening channel is stable at 10~15mg / m³, both of which are lower than the preset threshold. The dust collection rate is 99.85%, and there is no dust overflow, which meets the environmental protection control requirements.

[0077] 4. Production efficiency and labor costs: Production efficiency increased from 12t / h in the traditional process to 14.2t / h, an increase of 18.3%; no need for real-time manual inspection, reducing labor input by 2 people / shift, and reducing labor costs by more than 30%; abnormal situations were handled in a timely manner, and there were no production interruptions, significantly improving production continuity.

[0078] 5. Resource utilization rate: The secondary crushing and recovery rate of unqualified coarse particles is 85%, and the dust recovery rate is 100%. The resource utilization rate is 10.3% higher than that of traditional processes, which further reduces production costs and improves the economic benefits of enterprises.

[0079] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," "third," and similar terms used in this application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. The terms "an" or "a" and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms "comprising" or "including" and similar terms mean that the elements or objects preceding "comprising" or "including" encompass the elements or objects listed following "comprising" or "including" and their equivalents, and do not exclude other elements or objects. "Above," "below," "left," "right," etc., are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0080] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for air-based screening of dry sand making towers based on AI recognition, characterized in that, include: S1. Obtain sand information during the wind screening process; S2. Use the AI ​​recognition module to identify and analyze the sand information; S3. Adjust the sand-making parameters of the dry sand-making tower based on the analysis results.

2. The method for air screening of dry sand making towers based on AI recognition according to claim 1, characterized in that: The sand information includes images of the sand in the wind-powered screening channel, particle size, and dust concentration data.

3. The method for air screening of dry sand making towers based on AI recognition according to claim 1, characterized in that: Before the AI ​​recognition module in S2 identifies and analyzes the sand information, it is first trained. The training of the AI ​​recognition module includes: S21, acquiring stone samples with different raw material characteristics, different crushing particle sizes, and different dust contents during the dry sand making process, as well as individual and mixed samples of qualified sand and gravel aggregates, unqualified fine particles, unqualified coarse particles, and dust; S22, collecting image data, particle size data, and dust concentration data of the samples; S23, preprocessing and labeling the sample data, and inputting the preprocessed sample data into a deep convolutional neural network to train an AI recognition module with particle type recognition, particle size detection, and dust concentration recognition capabilities.

4. The method for air screening of dry sand making towers based on AI recognition according to claim 3, characterized in that: The sample data in S23 also includes crushing sample data under different wear levels of sand making equipment and screening sample data under different wind parameters.

5. The method for air screening of dry sand making towers based on AI recognition according to claim 1, characterized in that: S3. Adjusting the sand-making parameters of the dry sand-making plant based on the analysis results includes: determining whether the sand-making parameters of the dry sand-making plant need to be adjusted based on the analysis results; if it is determined that adjustment is needed, then the sand-making parameters of the dry sand-making plant are adjusted; if it is determined that adjustment is not needed, then the current sand-making parameters of the dry sand-making plant are kept unchanged.

6. The method for air screening of dry sand making towers based on AI recognition according to claim 4, characterized in that: When it is determined that adjustment is needed, the sand making parameters of the dry sand making tower are adjusted, and the wind screening method further includes: obtaining sand information during the wind screening process after the sand making parameters are adjusted; repeating S2 and S3.

7. The method for air screening of dry sand making towers based on AI recognition according to claim 1, characterized in that: S1 also includes acquiring the equipment status and operation data information of the dry sand making plant; S2 also includes using an AI recognition module to identify and analyze the equipment status and operation data information of the dry sand making plant.

8. The method for air screening of dry sand making towers based on AI recognition according to claim 1, characterized in that: The sand making parameters include the wind speed and air volume of the wind screening, and the crushing parameters of the sand making equipment.