Preparation system and method of health-preserving phosphorite sand based on cloud computing platform and DEEPSEEK

Abstract

The invention discloses a preparation system and method of health-preserving phosphorite sand based on a cloud computing platform and DEEPSEEK, belongs to the technical field of production of health-preserving phosphorite sand, and realizes three core progresses through deep learning collaborative optimization of the cloud computing platform and DEEPSEEK: firstly, the proportion of phosphorite sand and medical stone and the particle size are accurately controlled, the stability of the health-preserving effect of a product is ensured, and the stability of the health-preserving effect of the product is ensured; the qualified rate reaches 99%; secondly, grade fluctuation, resource utilization rate, energy consumption and pollution emission are synchronously optimized by adopting a multi-target dynamic weight algorithm, the resource utilization rate is greater than or equal to 85%, and the energy consumption is reduced by 20%; and thirdly, a real-time feedback mechanism is constructed, quality fluctuation is predicted through a Transform model, equipment is regulated and controlled in a closed loop mode, the abnormal response time is shortened from 30 minutes to 10 seconds, and green continuous production is supported.

Description

Technical Field

[0001] The present invention relates to the technical field of health-preserving phosphate ore production, and in particular to a system and method for preparing health-preserving phosphate ore based on a cloud computing platform and DEEPSEEK. Background Art

[0002] Phosphate ore is widely used in the health industry because it combines the mineral release properties of phosphate ore with the adsorption and health benefits of medical stone. The traditional preparation process relies on manual experience to control the raw material ratio and crushing and grinding parameters, which has significant defects: Low ratio accuracy: manual measurement error reaches ±5%, resulting in unstable health effects of the product; Rough particle size control: Grinding equipment relies on fixed parameters, and particle size deviation exceeds ±0.3 mm, affecting product application performance; Serious waste of resources: raw material utilization rate is less than 80%, energy consumption is high, and pollution emissions are large (such as phosphogypsum accumulation and excessive dust); Response lag: Process anomalies require manual intervention, and the response time exceeds 30 minutes, which cannot support continuous production.

[0003] Existing technologies mostly use basic automated control systems to control grinding parameters, but do not solve the problem of multi-objective optimization; some technologies also propose solutions for cloud computing platforms to manage production data, but lack the ability of deep learning models to predict quality fluctuations.

[0004] Therefore, there is an urgent need for an intelligent solution that integrates real-time prediction, multi-objective optimization and closed-loop control. Summary of the Invention

[0005] The purpose of the present invention is to provide a system and method for preparing healthy phosphate sand based on a cloud computing platform and DEEPSEEK to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above object, the present invention provides the following solutions: The present invention provides a system for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK, comprising a cloud computing platform, wherein the cloud computing platform integrates: Data acquisition module, used to collect raw material composition, process parameters, equipment status and environmental monitoring data through sensors; Data storage module, used to store data using distributed file systems, columnar databases, and time series databases; Data processing module, used to perform large-scale parallel computing, real-time stream processing and data mining; The model building module is used to build a production ratio model based on machine learning, which includes: Production ratio optimization algorithm: ; Among them, P is the comprehensive score of the product's health benefits, is the mixing ratio of phosphate sand and medical stone, α and β are weight coefficients trained through historical data, is the contribution of the two raw materials to the efficacy, and ε is the error correction term; Particle size PID control algorithm: ; in, Adjust parameters for grinding equipment, is the deviation between the real-time particle size and the target value, are proportional, integral, and differential coefficients, respectively, obtained through experiments or historical data optimization; Multi-objective dynamic weight algorithm: ; in, is the grade fluctuation value of finished phosphate sand, with a target range of ±0.5%∽1.0%. To achieve the goals in the production process Content percentage, It is the relative deviation rate between the actual finished product grade and the standard value, and is used to quantify the degree of fluctuation in the finished product quality. Comprehensive utilization rate of resources, required to be ≥85%, is the energy cost per unit output, is the pollutant emission, is the dynamic weight coefficient, which is adjusted by the DEEPSEEK model according to the real-time working conditions; The product quality prediction module is used to build a multimodal time series Transformer model using DEEPSEEK to predict P2O5 content and fluctuation range; Intelligent control module, used to dynamically adjust equipment parameters based on prediction results and execute resource scheduling optimization algorithms: ; in, is the energy consumption of the i-th device, The raw material waste rate.

[0007] Preferably, the cloud computing platform adopts a layered architecture, including an infrastructure layer, a virtualization layer, a platform service layer and an application layer.

[0008] Preferably, the data acquisition module uses an Internet of Things sensor to monitor the concentration of phosphogypsum in the tailings pond and the PM2.5 value of dust in real time and encode them in binary form according to the exceeding threshold value.

[0009] Preferably, the multimodal temporal Transformer model includes: The input embedding layer encodes heterogeneous information such as ore properties, process parameters, and environmental data into a unified vector: ; in, Ore characteristics include Content, particle size distribution, humidity, are process parameters, including crusher power, flotation reagent dosage, and drying temperature. Environmental monitoring data, including tailings pond phosphogypsum concentration and dust PM2.5 value, are binary coded based on the exceeding threshold. is a trainable embedding weight matrix; Multi-head self-attention mechanism calculates the dynamic correlation weights between features: ; Among them, Attention is attention, Softmax is the normalization function, which is used to Converted to probability distribution, Q, K, V are query, key, and value matrices respectively, generated by linear transformation of input embedding E, is the attention head dimension, with a value range of 64-256, which controls the smoothness of the weight distribution. The output is to capture the nonlinear relationship between phosphate rock characteristics and process parameters; Prediction output layer, predicting finished products Content and quality fluctuation range: ; ; in, For the next production cycle of finished phosphate sand The predicted value of the content, MLP is multi-layer perception, the output dimension is 1, The predicted value of is the output of the L-th layer Transformer, ranging from 6 to 12 layers, is the error measure of the dependent variable y, The quality fluctuation error is required to be ≤0.5%. is the predicted value corresponding to the i-th observation value, is the actual value of the ith observation.

[0010] Preferably, the intelligent control module further includes: Using dynamic parameter adjustment algorithm and multi-objective optimization function, process parameter adjustment instructions are generated based on the prediction results: ; in, Target Standard value, is the energy cost, is the pollution penalty term, It is a dynamic weight value, which is calculated in real time by the DEEPSEEK-MoE architecture and has a value range of ; Real-time feedback learning mechanism, the loss function combines prediction error and process constraints for online training: ; in, is the loss amount, MSE is the mean square error, and weight =0.7, ReLU is the rectified linear unit, which is the constraint violation penalty term, is the resource weight, For environmental protection weight, Comprehensive utilization rate of resources, required to be ≥85%, The model parameters are updated every hour through the cloud platform based on the pollutant emissions.

[0011] The present invention also provides a method for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK, comprising the following steps: S1: Raw material screening and pretreatment: Use a vibrating screen to remove oversized particles in phosphate sand and medical stone, and perform component analysis on the medical stone; S2: Mixing ratio control: The optimal ratio is calculated through the cloud computing platform, and the amount of raw materials added is adjusted in real time using intelligent metering equipment; S3: Crushing and grinding optimization: Dynamically adjust crusher power, grinding speed and feed rate based on the DEEPSEEK prediction model; S4: Grading and screening feedback: Use a laser particle size analyzer to detect particle size and feed it back to the cloud computing platform for closed-loop control of subsequent equipment.

[0012] Preferably, in step S3, the prediction model construction includes: S31. The ore P2O5 content, humidity, crusher power, and flotation reagent dosage are used as input features; S32. Output the predicted value and fluctuation error of P2O5 content through the multimodal time series Transformer model.

[0013] Preferably, in step S4, the hierarchical screening is specifically as follows: S41. Set up a multi-stage vibrating screening device to classify the mixed material by particle size; S42. When the particle size deviates from the target, the grinder feed speed is automatically reduced and the grinding time is extended.

[0014] Compared with the prior art, the present invention has achieved the following beneficial technical effects: 1. The present invention adopts a cloud computing platform to achieve remote monitoring and data sharing, improving the convenience and real-time performance of production management; 2. The present invention integrates DEEPSEEK deep learning technology to achieve intelligent control and optimization of the manufacturing process, improving production efficiency and product quality; 3. The present invention realizes collaborative operation and intelligent management among devices, thus improving equipment utilization and production efficiency; 4. The present invention ensures that the particle size of phosphate sand particles is stable within the range of 1 mm or 2 mm through precise particle size control technology and data processing, meeting specific application requirements; 5. The present invention accurately controls the mixing ratio of phosphate sand and medical stone, giving full play to the health-preserving effects of both and improving product quality; 6. The present invention provides technical support for the sustainable development of the health-preserving phosphate sand manufacturing industry through data analysis and intelligent optimization. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] The purpose of the present invention is to provide a system and method for preparing healthy phosphate sand based on a cloud computing platform and DEEPSEEK to solve the problems existing in the prior art.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in conjunction with specific embodiments.

[0018] Example 1 This embodiment provides a system for preparing healthy phosphate sand based on a cloud computing platform and DEEPSEEK. The system includes a cloud computing platform with a layered architecture design, including an infrastructure layer (computing, storage, and network resources), a virtualization layer (abstracting and encapsulating physical resources), a cloud computing platform layer (providing services and tools), and an application layer (deploying related applications). The cloud computing platform integrates: The data acquisition module is used to collect raw material, process, and equipment data during the preparation process through various sensors and monitoring equipment, such as raw material composition, production process parameters, equipment operating status, and product quality data. It is then binary-coded using exceeding thresholds and transmitted to the cloud computing platform in real time. The data acquisition module uses IoT sensors and IoT monitoring equipment. Data storage module, used to store data using distributed file systems, columnar databases, and time series databases; The data storage solution can be: Distributed file system: Choose a distributed file system such as Ceph to store data in multiple independent nodes and ensure data reliability through multi-copy storage; Column-based storage databases: For structured data, such as device parameters, column-based storage databases (such as Apache HBase) are used to store them to improve query performance. Time series databases, such as InfluxDB, are used for time series data, such as data collected by sensors, which support high-speed writing and querying. Object storage: For unstructured data such as product images and videos, object storage (such as Amazon S3) is used, which has the advantages of high scalability and low cost; Data processing module, used to perform large-scale parallel computing, real-time stream processing and data mining; Data processing includes: Large-scale parallel computing. Cloud computing platforms have large-scale parallel computing capabilities, which can decompose computing tasks and distribute them to multiple computing nodes for parallel execution, shortening processing time. For example, when analyzing large amounts of historical data, operations can be completed quickly. Real-time stream processing: supports real-time stream processing, receiving and processing continuous data streams generated by sensors and devices in real time, and promptly detecting anomalies and issuing warnings; for example, immediate alarms when device parameters are out of range; Data mining and machine learning: integrating data mining and machine learning algorithms to extract information and patterns from massive amounts of data; for example, establishing predictive models and optimizing production processes; Data visualization: provides data visualization function, displays data in intuitive charts, and facilitates users to understand and analyze production data; Remote monitoring, using big data processing and visualization technology to achieve remote real-time monitoring of the entire manufacturing process; Analysis and mining: using the computing power of cloud computing platforms and adopting data mining, machine learning and other technologies to conduct in-depth analysis of data; The model building module is used to build a production ratio model based on machine learning, which includes: Production ratio optimization algorithm: ; Among them, P is the comprehensive score of the product's health benefits. is the mixing ratio of phosphate sand and medical stone, α and β are weight coefficients trained through historical data, is the contribution of the two raw materials to the efficacy, and ε is the error correction term; Particle size PID control algorithm: ; in, Adjust parameters for grinding equipment, is the deviation between the real-time particle size and the target value, are proportional, integral, and differential coefficients, respectively, obtained through experiments or historical data optimization; Multi-objective dynamic weight algorithm: ; in, is the grade fluctuation value of finished phosphate sand, with a target range of ±0.5%∽1.0%. To achieve the goals in the production process Content percentage, It is the relative deviation rate between the actual finished product grade and the standard value, and is used to quantify the degree of fluctuation in the finished product quality. Comprehensive utilization rate of resources, required to be ≥85%, is the energy cost per unit output, is the amount of pollutant emissions (such as phosphogypsum, dust), is the dynamic weight coefficient, which is adjusted by the DEEPSEEK model according to the real-time working conditions; The product quality prediction module uses DEEPSEEK to build a multimodal time series Transformer model to predict P2O5 content and fluctuation range. The multimodal time series Transformer model includes: The input embedding layer encodes heterogeneous information such as ore properties, process parameters, and environmental data into a unified vector: ; in, Ore characteristics include Content, particle size distribution, humidity, are process parameters, including crusher power, flotation reagent dosage, and drying temperature. Environmental monitoring data, including tailings pond phosphogypsum concentration and dust PM2.5 value, are binary coded based on the exceeding threshold. is a trainable embedding weight matrix; Multi-head self-attention mechanism calculates the dynamic correlation weights between features: ; Among them, Attention is attention, Softmax is the normalization function, which is used to Converted to probability distribution, Q, K, V are query, key, and value matrices respectively, generated by linear transformation of input embedding E, is the attention head dimension, with a value range of 64-256, which controls the smoothness of the weight distribution. The output is to capture the nonlinear relationship between phosphate rock characteristics and process parameters; Prediction output layer, predicting finished products Content and quality fluctuation range: ; ; in, For the next production cycle of finished phosphate sand The predicted value of the content, MLP is multi-layer perception, the output dimension is 1, The predicted value of is the output of the L-th layer Transformer, ranging from 6 to 12 layers, is the error measure of the dependent variable y, The quality fluctuation error is required to be ≤0.5%. is the predicted value corresponding to the i-th observation value, is the actual value of the ith observation.

[0019] Intelligent control module, used to dynamically adjust equipment parameters based on prediction results and execute resource scheduling optimization algorithms: ; in, is the energy consumption of the i-th device, is the raw material waste rate; The intelligent control module also includes: Using dynamic parameter adjustment algorithm and multi-objective optimization function, process parameter adjustment instructions are generated based on the prediction results: ; in, Target Standard value, is the energy cost, is the pollution penalty term, It is a dynamic weight value, which is calculated in real time by the DEEPSEEK-MoE architecture and has a value range of ; Real-time feedback learning mechanism, the loss function combines prediction error and process constraints for online training: ; in, is the loss amount, MSE is the mean square error, and weight =0.7, ReLU is the rectified linear unit, which is the constraint violation penalty term, is the resource weight, For environmental protection weight, Comprehensive utilization rate of resources, required to be ≥85%, The model parameters are updated every hour through the cloud platform based on the pollutant emissions.

[0020] Example 2: A method for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK, comprising the following steps: S1: Raw material screening and pretreatment: Use a vibrating screen to remove oversized particles in phosphate sand and medical stone, and perform component analysis on the medical stone; S2: Mixing ratio control: The optimal ratio is calculated through the cloud computing platform, and the amount of raw materials added is adjusted in real time using intelligent metering equipment; S3: Crushing and Grinding Optimization: Dynamically adjust crusher power, grinding speed, and feed rate based on the DEEPSEEK prediction model; prediction model construction includes: S31. The ore P2O5 content, humidity, crusher power, and flotation reagent dosage are used as input features; S32. Output the predicted P2O5 content and fluctuation error through the multimodal time series Transformer model; S4: Grading and screening feedback: Use a laser particle size analyzer to detect particle size and provide feedback to the cloud computing platform for closed-loop control of subsequent equipment; Grading and screening specifically include: S41. Set up a multi-stage vibrating screening device to classify the mixed material by particle size; S42. When the particle size deviates from the target, the grinder feed speed is automatically reduced and the grinding time is extended.

[0021] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0022] It should be noted that the components mentioned in the above embodiments are all universal standard parts or components known to those skilled in the art, and their structures and principles can be known to those skilled in the art through technical manuals or conventional experimental methods.

[0023] The present invention uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A system for preparing healthy phosphate sand based on a cloud computing platform and DEEPSEEK, comprising a cloud computing platform, characterized in that: The cloud computing platform integrates: Data acquisition module, used to collect raw material composition, process parameters, equipment status and environmental monitoring data through sensors; Data storage module, used to store data using distributed file systems, columnar databases, and time series databases; Data processing module, used to perform large-scale parallel computing, real-time stream processing and data mining; The model building module is used to build a production ratio model based on machine learning, which includes: Production ratio optimization algorithm: ; Among them, P is the comprehensive score of the product's health benefits, is the mixing ratio of phosphate sand and medical stone, α and β are weight coefficients trained through historical data, is the contribution of the two raw materials to the efficacy, and ε is the error correction term; Particle size PID control algorithm: ; in, Adjust parameters for grinding equipment, is the deviation between the real-time particle size and the target value, are proportional, integral, and differential coefficients, respectively, obtained through experiments or historical data optimization; Multi-objective dynamic weight algorithm: ; in, is the grade fluctuation value of finished phosphate sand, with a target range of ±0.5%∽1.0%. To achieve the goals in the production process Content percentage, It is the relative deviation rate between the actual finished product grade and the standard value, and is used to quantify the degree of fluctuation in the finished product quality. Comprehensive utilization rate of resources, required to be ≥85%, is the energy cost per unit output, is the pollutant emission, is the dynamic weight coefficient, which is adjusted by the DEEPSEEK model according to the real-time working conditions; The product quality prediction module is used to build a multimodal time series Transformer model using DEEPSEEK to predict P2O5 content and fluctuation range; Intelligent control module, used to dynamically adjust equipment parameters based on prediction results and execute resource scheduling optimization algorithms: ; in, is the energy consumption of the i-th device, The raw material waste rate.

2. The system for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK according to claim 1, characterized in that: The cloud computing platform adopts a layered architecture, including an infrastructure layer, a virtualization layer, a platform service layer and an application layer.

3. The system for preparing health-preserving phosphate sand based on cloud computing platform and DEEPSEEK according to claim 1, characterized in that: The data acquisition module uses Internet of Things sensors to monitor the concentration of phosphogypsum in the tailings pond and the PM2.5 value of dust in real time and encode them in binary form according to the exceeding threshold value.

4. The system for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK according to claim 1, characterized in that: The multimodal temporal Transformer model includes: The input embedding layer encodes heterogeneous information such as ore properties, process parameters, and environmental data into a unified vector: ; in, Ore characteristics include Content, particle size distribution, humidity, are process parameters, including crusher power, flotation reagent dosage, and drying temperature. Environmental monitoring data, including tailings pond phosphogypsum concentration and dust PM2.5 value, are binary coded based on the exceeding threshold. is a trainable embedding weight matrix; Multi-head self-attention mechanism calculates the dynamic correlation weights between features: ; Among them, Attention is attention, Softmax is the normalization function, which is used to Converted to probability distribution, Q, K, V are query, key, and value matrices respectively, generated by linear transformation of input embedding E, is the attention head dimension, with a value range of 64-256, which controls the smoothness of the weight distribution. The output is to capture the nonlinear relationship between phosphate rock characteristics and process parameters; Prediction output layer, predicting finished products Content and quality fluctuation range: ; ; in, For the next production cycle of finished phosphate sand The predicted value of the content, MLP is multi-layer perception, the output dimension is 1, The predicted value of is the output of the L-th layer Transformer, ranging from 6 to 12 layers, is the error measure of the dependent variable y, The quality fluctuation error is required to be ≤0.5%. is the predicted value corresponding to the i-th observation value, is the actual value of the ith observation.

5. The system for preparing health-preserving phosphate sand based on cloud computing platform and DEEPSEEK according to claim 1, characterized in that: The intelligent control module further includes: Using dynamic parameter adjustment algorithm and multi-objective optimization function, process parameter adjustment instructions are generated based on the prediction results: ; in, Target Standard value, is the energy cost, is the pollution penalty term, It is a dynamic weight value, which is calculated in real time by the DEEPSEEK-MoE architecture and has a value range of ; Real-time feedback learning mechanism, the loss function combines prediction error and process constraints for online training: ; in, is the loss amount, MSE is the mean square error, and weight =0.7, ReLU is the rectified linear unit, which is the constraint violation penalty term, is the resource weight, For environmental protection weight, Comprehensive utilization rate of resources, required to be ≥85%, The model parameters are updated every hour through the cloud platform based on the pollutant emissions.

6. A method for preparing health-preserving phosphate sand based on a cloud computing platform and DEEPSEEK, characterized by: The following steps are involved: S1: Raw material screening and pretreatment: Use a vibrating screen to remove oversized particles in phosphate sand and medical stone, and perform component analysis on the medical stone; S2: Mixing ratio control: The optimal ratio is calculated through the cloud computing platform, and the amount of raw materials added is adjusted in real time using intelligent metering equipment; S3: Crushing and grinding optimization: Dynamically adjust crusher power, grinding speed and feed rate based on the DEEPSEEK prediction model; S4: Grading and screening feedback: Use a laser particle size analyzer to detect particle size and feed it back to the cloud computing platform for closed-loop control of subsequent equipment.

7. The method for preparing health-preserving phosphate sand based on cloud computing platform and DEEPSEEK according to claim 6, characterized in that: In step S3, the prediction model construction includes: S31. The ore P2O5 content, humidity, crusher power, and flotation reagent dosage are used as input features; S32. Output the predicted value and fluctuation error of P2O5 content through the multimodal time series Transformer model.

8. The method for preparing health-preserving phosphate sand based on cloud computing platform and DEEPSEEK according to claim 1, characterized in that: In step S4, the hierarchical screening is specifically as follows: S41. Set up a multi-stage vibrating screening device to classify the mixed material by particle size; S42. When the particle size deviates from the target, the grinder feed speed is automatically reduced and the grinding time is extended.