A system and method for controlling the particle size distribution of ultrafine high-purity spherical silicon powder

Abstract

The application discloses a kind of superfine high-purity spherical silicon micro-powder particle size distribution regulation and control system and method, belong to spherical silicon micro-powder technical field, the system includes: process parameter acquisition module is configured to collect superfine high-purity spherical silicon micro-powder preparation process real-time data;Process parameter processing module is configured to process the collected superfine high-purity spherical silicon micro-powder preparation process real-time data;Particle size distribution regulation and control module is configured to build superfine high-purity spherical silicon micro-powder particle size distribution prediction model, predict superfine high-purity spherical silicon micro-powder particle size distribution situation, and carry out closed-loop regulation and control to superfine high-purity spherical silicon micro-powder particle size distribution.The application solves the problem that existing superfine high-purity spherical silicon micro-powder particle size distribution cannot be effectively regulated and controlled, and cannot better meet the use requirements.The application can effectively regulate and control superfine high-purity spherical silicon micro-powder particle size distribution, improve the use effect of prepared superfine high-purity spherical silicon micro-powder, and can better meet the use requirements.

Description

Technical Field

[0001] This invention relates to the field of spherical silicon micropowder technology, specifically to a system and method for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder. Background Technology

[0002] Spherical silicon micropowder, with its high purity, fine particle characteristics, excellent dielectric properties and thermal conductivity, as well as its low coefficient of expansion, plays an irreplaceable and important role in many fields such as integrated circuit packaging, aerospace, coating manufacturing, pharmaceutical production and daily cosmetics.

[0003] Chinese patent application CN109455726A discloses a method for preparing spherical silicon micropowder for electronic packaging. The method involves using a flame combustion method to prepare spherical silicon micropowder of different particle sizes from angular silicon micropowder. The spherical silicon micropowder product then undergoes coarse and fine classification processes, followed by particle size distribution. These processes remove large particles, controlling the content of large particles to below 10, effectively controlling large particles and preventing them from clogging the encapsulation injection port or causing poor encapsulation during downstream product use. However, this patent has the following drawbacks:

[0004] Existing technologies cannot effectively control the particle size distribution of ultrafine high-purity spherical silicon micropowder, resulting in poor performance of the prepared ultrafine high-purity spherical silicon micropowder, which cannot meet the application requirements. Summary of the Invention

[0005] The purpose of this invention is to provide a system and method for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder, which can effectively control the particle size distribution of ultrafine high-purity spherical silicon micropowder, improve the performance of the prepared ultrafine high-purity spherical silicon micropowder, better meet the application requirements, and solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A system for controlling the particle size distribution of ultrafine, high-purity spherical silica powder includes:

[0008] The process parameter acquisition module is configured to acquire real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder.

[0009] The process parameter processing module is configured to process the real-time data of the collected ultrafine high-purity spherical silicon micropowder preparation process and determine the characteristic data of the ultrafine high-purity spherical silicon micropowder preparation process.

[0010] The particle size distribution control module is configured to construct a particle size distribution prediction model for ultrafine high-purity spherical silicon micropowder, analyze the process characteristic data of ultrafine high-purity spherical silicon micropowder preparation, predict the particle size distribution of ultrafine high-purity spherical silicon micropowder, and perform closed-loop control on the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0011] Preferably, real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder is collected, including:

[0012] Real-time monitoring and data acquisition were conducted on the raw material particle size, melting temperature, melting time, cooling rate, airflow classification speed, airflow pressure, classifier rotation speed, sieve mesh size, and vibration frequency during the preparation of ultrafine high-purity spherical silicon micropowder to determine the real-time data of the ultrafine high-purity spherical silicon micropowder preparation process.

[0013] Preferably, the real-time data processing of the preparation process of ultrafine high-purity spherical silicon micropowder includes:

[0014] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is cleaned to remove noise that is useless for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0015] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder was checked to identify duplicate, missing and outlier values ​​in the real-time data, and the duplicate, missing and outlier values ​​in the real-time data were processed.

[0016] Specifically, duplicate values ​​in the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder are removed, and it is determined whether the missing and outlier values ​​in the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder are useful for the control of particle size distribution of ultrafine high-purity spherical silicon micropowder. If they are useful, the missing values ​​are filled with the mean and the outlier values ​​are replaced with the median. If they are not useful, the missing and outlier values ​​are directly deleted.

[0017] Preferably, the real-time data processing for the preparation process of ultrafine high-purity spherical silica powder also includes:

[0018] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is normalized to convert it into a unified data format, remove the dimensional differences in the real-time data, and determine the standardized real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder.

[0019] Feature extraction was performed on real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder. Features useful for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder were extracted from the real-time data of the preparation process, and the feature data of the preparation process of ultrafine high-purity spherical silicon micropowder were determined.

[0020] Preferably, a particle size distribution prediction model for ultrafine, high-purity spherical silica powder is constructed, including:

[0021] Based on the requirements for particle size distribution control of ultrafine high-purity spherical silicon micropowder, historical data on the preparation process of ultrafine high-purity spherical silicon micropowder, including the particle size distribution of ultrafine high-purity spherical silicon micropowder, were collected.

[0022] The collected historical data on the preparation process of ultrafine high-purity spherical silicon micropowder was divided into training set and test set.

[0023] Based on deep learning technology, a training set is used to train the deep learning model, enabling the deep learning model to autonomously learn the relationship between the particle size distribution of ultrafine high-purity spherical silicon micropowder and the preparation process of ultrafine high-purity spherical silicon micropowder, and automatically learn the particle size distribution prediction behavior of ultrafine high-purity spherical silicon micropowder, thus determining the particle size distribution prediction model of ultrafine high-purity spherical silicon micropowder based on deep learning.

[0024] The performance of the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon micropowder was tested using a test set to evaluate whether the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon micropowder can achieve the effect of predicting the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0025] When the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon powder fails to achieve the desired effect, the parameters of the deep learning-based prediction model are adjusted, and the model is continuously optimized until it can achieve the desired effect. The optimal prediction model for the particle size distribution of ultrafine high-purity spherical silicon powder is then determined.

[0026] Preferably, the process characteristic data of ultrafine high-purity spherical silicon micropowder preparation are analyzed to predict the particle size distribution of ultrafine high-purity spherical silicon micropowder, including:

[0027] Obtain the optimal particle size distribution prediction model for ultrafine high-purity spherical silicon powder, and deploy the optimal particle size distribution prediction model for ultrafine high-purity spherical silicon powder in the actual particle size distribution prediction environment.

[0028] The process characteristics data of ultrafine high-purity spherical silicon powder preparation are input into the optimal particle size distribution prediction model of ultrafine high-purity spherical silicon powder. Based on the optimal particle size distribution prediction model of ultrafine high-purity spherical silicon powder preparation process characteristics data are analyzed, and the particle size distribution of ultrafine high-purity spherical silicon powder is predicted, thus determining the particle size distribution prediction result of ultrafine high-purity spherical silicon powder.

[0029] Preferably, closed-loop control is performed on the particle size distribution of ultrafine, high-purity spherical silica powder, including:

[0030] The predicted particle size distribution of ultrafine high-purity spherical silicon powder was compared and analyzed with the pre-set particle size distribution threshold of ultrafine high-purity spherical silicon powder.

[0031] When the predicted particle size distribution of ultrafine high-purity spherical silicon powder is not within the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder, the particle size distribution of ultrafine high-purity spherical silicon powder is adjusted.

[0032] The process involves adjusting the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder, and monitoring the particle size distribution of the ultrafine high-purity spherical silicon micropowder after the process adjustment. Based on the monitoring results, the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is adjusted and optimized to form a closed-loop control of the particle size distribution of ultrafine high-purity spherical silicon micropowder until the particle size distribution of the prepared ultrafine high-purity spherical silicon micropowder is uniform.

[0033] Preferably, the real-time data for the preparation process of ultrafine high-purity spherical silicon micropowder is adjusted, including:

[0034] When the predicted particle size distribution of ultrafine high-purity spherical silicon powder is not within the threshold range of ultrafine high-purity spherical silicon powder particle size distribution, the deviation value between the predicted particle size distribution of ultrafine high-purity spherical silicon powder and the threshold range of ultrafine high-purity spherical silicon powder particle size distribution is extracted.

[0035] The deviation value is compared with the threshold range of particle size distribution of ultrafine high-purity spherical silicon micropowder to obtain the deviation ratio value.

[0036] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder was adjusted according to the gradient adjustment based on the initial process real-time data;

[0037] Extract the deviation between the predicted particle size distribution of ultrafine high-purity spherical silicon micropowder after each adjustment and the threshold range of particle size distribution of ultrafine high-purity spherical silicon micropowder;

[0038] The dynamic deviation ratio is obtained by comparing the deviation between the predicted particle size distribution of ultrafine high-purity spherical silicon powder after each adjustment and the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder with the span of the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder.

[0039] Compare the dynamic deviation ratio value with the deviation ratio value;

[0040] When the dynamic deviation ratio exceeds the deviation ratio, the real-time process data adjustment gradient is adjusted to reduce the real-time process data adjustment gradient, preventing excessive fluctuations in the real-time process data adjustment that could lead to poor control stability and imbalance.

[0041] Preferably, when the dynamic deviation ratio exceeds the deviation ratio, the adjustment gradient of the real-time process data is adjusted, including:

[0042] When the dynamic deviation ratio value exceeds the deviation ratio value, obtain the ratio deviation value between the dynamic deviation ratio value and the deviation ratio value;

[0043] The real-time reaction temperature during the preparation of ultrafine high-purity spherical silicon micropowder is retrieved.

[0044] The real-time reaction temperature in the current preparation process of ultrafine high-purity spherical silicon micropowder is compared with the reference temperature to obtain the temperature deviation value;

[0045] The real-time gas pressure in the reaction chamber during the current preparation process of ultrafine high-purity spherical silicon micropowder was obtained;

[0046] The gas pressure in the real-time reaction chamber during the current preparation process of ultrafine high-purity spherical silicon micropowder is compared with the reference gas pressure to obtain the gas pressure deviation value.

[0047] The temperature deviation value and the gas pressure deviation value are standardized to obtain the standardized temperature deviation value and gas pressure deviation value.

[0048] The gradient adjustment coefficient is obtained by combining the standardized temperature deviation value and gas pressure deviation value with the proportional deviation value between the dynamic deviation ratio value and the deviation ratio value.

[0049] The gradient adjustment coefficient is obtained by the following formula:

[0050]

[0051] Where δ represents the gradient adjustment coefficient; P t This represents the temperature deviation value after standardization; P a B represents the gas pressure deviation value after standardization; D represents the dynamic deviation ratio value and the deviation ratio value;

[0052] The gradient adjustment coefficient is used to adjust the real-time process data adjustment gradient, wherein the adjusted real-time process data adjustment gradient is obtained by the following formula:

[0053]

[0054] Among them, G t G0 represents the adjusted real-time process data adjustment gradient; δ represents the gradient adjustment coefficient; D b The value represents the standard deviation of the dynamic deviation ratio; e represents the preset minimum coefficient to prevent ineffective gradient adjustment when the standard deviation of the dynamic deviation ratio is 0.

[0055] According to another aspect of the present invention, a method for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder is provided, which is based on the above-mentioned ultrafine high-purity spherical silicon micropowder particle size distribution control system, and includes the following steps:

[0056] Real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder was collected and processed to determine the characteristic data of the preparation process of ultrafine high-purity spherical silicon micropowder.

[0057] A particle size distribution prediction model for ultrafine high-purity spherical silicon micropowder was constructed. The characteristic data of the preparation process of ultrafine high-purity spherical silicon micropowder were analyzed to determine the predicted particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0058] The predicted particle size distribution of ultrafine high-purity spherical silicon powder was analyzed, and the particle size distribution of ultrafine high-purity spherical silicon powder was controlled in a closed loop according to the analysis until the particle size distribution of the prepared ultrafine high-purity spherical silicon powder was uniform.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] This invention achieves real-time monitoring and data acquisition of raw material particle size, melting temperature, melting time, cooling rate, airflow classification speed, airflow pressure, classifier rotation speed, sieve mesh size, and vibration frequency during the preparation of ultrafine high-purity spherical silicon micropowder. This allows for the determination of real-time process data for ultrafine high-purity spherical silicon micropowder preparation. Furthermore, this real-time data is processed to determine the characteristic data of the ultrafine high-purity spherical silicon micropowder preparation process. Based on the particle size distribution control requirements of ultrafine high-purity spherical silicon micropowder, a method for constructing ultrafine high-purity spherical silicon micropowder preparation is developed. A particle size distribution prediction model is used to analyze the process characteristic data of ultrafine high-purity spherical silicon micropowder preparation, determine the predicted particle size distribution of ultrafine high-purity spherical silicon micropowder, and analyze the predicted particle size distribution. Based on the analysis, the particle size distribution of ultrafine high-purity spherical silicon micropowder is controlled in a closed loop until the particle size distribution of the prepared ultrafine high-purity spherical silicon micropowder is uniform. This model can effectively control the particle size distribution of ultrafine high-purity spherical silicon micropowder, improve the application effect of the prepared ultrafine high-purity spherical silicon micropowder, and better meet the application requirements. Attached Figure Description

[0061] Figure 1 This is a block diagram of the ultrafine high-purity spherical silicon micropowder particle size distribution control system of the present invention;

[0062] Figure 2 This is a flowchart of the method for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder according to the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] To address the problem that existing technologies cannot effectively control the particle size distribution of ultrafine, high-purity spherical silica powder, resulting in poor performance and failure to meet application requirements, please refer to [the relevant documentation / reference]. Figures 1-2 This embodiment provides the following technical solution:

[0065] A particle size distribution control system for ultrafine high-purity spherical silicon micropowder includes: a process parameter acquisition module, a process parameter processing module, and a particle size distribution control module.

[0066] Specifically, through the interaction between the process parameter acquisition module, the process parameter processing module, and the particle size distribution control module, the particle size distribution of ultrafine high-purity spherical silicon micropowder can be effectively controlled, thereby improving the performance of the prepared ultrafine high-purity spherical silicon micropowder and better meeting the application requirements.

[0067] Among them, the process parameter acquisition module is used to collect real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder;

[0068] In this embodiment, real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder is collected, including:

[0069] Real-time monitoring and data acquisition were conducted on the raw material particle size, melting temperature, melting time, cooling rate, airflow classification speed, airflow pressure, classifier rotation speed, sieve mesh size, and vibration frequency during the preparation of ultrafine high-purity spherical silicon micropowder to determine the real-time data of the ultrafine high-purity spherical silicon micropowder preparation process.

[0070] It should be noted that: raw material particle size refers to the initial crushed size of the raw silica, which affects the subsequent melting and spheroidization effect; melting temperature refers to the temperature of the particles in the molten state. The melting temperature of silica powder directly affects the spheroidization effect of the particles. Too high a temperature may cause the particles to over-melt, while too low a temperature will prevent the formation of spherical shapes; melting time refers to the residence time of the particles in the molten state, which affects the integrity of spheroidization and particle size; cooling rate refers to the cooling speed of the molten particles, which affects the stability of the spherical structure and the crystallization state of the particles; air classifying speed refers to the airflow speed of the air classifier, which determines the particle classification effect and affects the uniformity of the final particle size distribution; airflow pressure refers to the airflow pressure in the air classifier, which affects the particle classification accuracy; classifier rotation speed refers to the rotation speed of the rotor inside the classifier, which affects the particle classification accuracy; and the screen aperture and vibration frequency of the screening machine determine the size range of the particles after screening.

[0071] The process parameter processing module is used to process the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder and determine the characteristic data of the preparation process of ultrafine high-purity spherical silicon micropowder.

[0072] In this embodiment, the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is processed, including:

[0073] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is cleaned to remove noise that is useless for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0074] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder was checked to identify duplicate, missing and outlier values ​​in the real-time data, and the duplicate, missing and outlier values ​​in the real-time data were processed.

[0075] Specifically, duplicate values ​​in the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder are removed, and it is determined whether the missing values ​​and outliers in the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder are useful for the control of particle size distribution of ultrafine high-purity spherical silicon micropowder. If they are useful, the missing values ​​are filled with the mean and the outliers are replaced with the median. If they are not useful, the missing values ​​and outliers are directly deleted.

[0076] It should be noted that cleaning the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder can improve the data quality.

[0077] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is normalized to convert it into a unified data format, remove the dimensional differences in the real-time data, and determine the standardized real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder.

[0078] Feature extraction was performed on real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder. Features useful for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder were extracted from the real-time data of the preparation process, and the feature data of the preparation process of ultrafine high-purity spherical silicon micropowder were determined.

[0079] The particle size distribution control module is used to construct a particle size distribution prediction model for ultrafine high-purity spherical silicon micropowder, analyze the process characteristic data of ultrafine high-purity spherical silicon micropowder preparation, predict the particle size distribution of ultrafine high-purity spherical silicon micropowder, and perform closed-loop control on the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0080] In this embodiment, a particle size distribution prediction model for ultrafine high-purity spherical silicon micropowder is constructed, including:

[0081] Based on the requirements for particle size distribution control of ultrafine high-purity spherical silicon micropowder, historical data on the preparation process of ultrafine high-purity spherical silicon micropowder, including the particle size distribution of ultrafine high-purity spherical silicon micropowder, were collected.

[0082] The collected historical data on the preparation process of ultrafine high-purity spherical silicon micropowder was divided into training set and test set.

[0083] Based on deep learning technology, a training set is used to train the deep learning model, enabling the deep learning model to autonomously learn the relationship between the particle size distribution of ultrafine high-purity spherical silicon micropowder and the preparation process of ultrafine high-purity spherical silicon micropowder, and automatically learn the particle size distribution prediction behavior of ultrafine high-purity spherical silicon micropowder, thus determining the particle size distribution prediction model of ultrafine high-purity spherical silicon micropowder based on deep learning.

[0084] The performance of the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon micropowder was tested using a test set to evaluate whether the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon micropowder can achieve the effect of predicting the particle size distribution of ultrafine high-purity spherical silicon micropowder.

[0085] When the deep learning-based prediction model for the particle size distribution of ultrafine high-purity spherical silicon powder fails to achieve the desired effect, the parameters of the deep learning-based prediction model are adjusted, and the model is continuously optimized until it can achieve the desired effect. The optimal prediction model for the particle size distribution of ultrafine high-purity spherical silicon powder is then determined.

[0086] In this embodiment, the characteristic data of the preparation process of ultrafine high-purity spherical silicon micropowder are analyzed to predict the particle size distribution of ultrafine high-purity spherical silicon micropowder, including:

[0087] Obtain the optimal particle size distribution prediction model for ultrafine high-purity spherical silicon powder, and deploy the optimal particle size distribution prediction model for ultrafine high-purity spherical silicon powder in the actual particle size distribution prediction environment.

[0088] The process characteristics data of ultrafine high-purity spherical silicon powder preparation are input into the optimal particle size distribution prediction model of ultrafine high-purity spherical silicon powder. Based on the optimal particle size distribution prediction model of ultrafine high-purity spherical silicon powder preparation process characteristics data are analyzed, and the particle size distribution of ultrafine high-purity spherical silicon powder is predicted, thus determining the particle size distribution prediction result of ultrafine high-purity spherical silicon powder.

[0089] In this embodiment, closed-loop control of the particle size distribution of ultrafine high-purity spherical silicon powder is performed, including:

[0090] The predicted particle size distribution of ultrafine high-purity spherical silicon powder was compared and analyzed with the pre-set particle size distribution threshold of ultrafine high-purity spherical silicon powder.

[0091] When the predicted particle size distribution of ultrafine high-purity spherical silicon powder is not within the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder, the particle size distribution of ultrafine high-purity spherical silicon powder is adjusted.

[0092] The process involves adjusting the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder, and monitoring the particle size distribution of the ultrafine high-purity spherical silicon micropowder after the process adjustment. Based on the monitoring results, the real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder is adjusted and optimized to form a closed-loop control of the particle size distribution of ultrafine high-purity spherical silicon micropowder until the particle size distribution of the prepared ultrafine high-purity spherical silicon micropowder is uniform.

[0093] To better illustrate the process of controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder, this embodiment provides a method for controlling the particle size distribution of ultrafine high-purity spherical silicon micropowder, based on the aforementioned particle size distribution control system for ultrafine high-purity spherical silicon micropowder, including the following steps:

[0094] Real-time data on the preparation process of ultrafine high-purity spherical silicon micropowder was collected and processed to determine the characteristic data of the preparation process. A particle size distribution prediction model for ultrafine high-purity spherical silicon micropowder was constructed, and the characteristic data of the preparation process were analyzed to determine the predicted particle size distribution. The predicted particle size distribution was analyzed, and the particle size distribution of ultrafine high-purity spherical silicon micropowder was controlled in a closed loop according to the analysis until the particle size distribution of the prepared ultrafine high-purity spherical silicon micropowder was uniform. This method can effectively control the particle size distribution of ultrafine high-purity spherical silicon micropowder, improve the application effect of the prepared ultrafine high-purity spherical silicon micropowder, and better meet the application requirements.

[0095] Specifically, adjustments were made to the real-time data of the preparation process for ultrafine, high-purity spherical silicon micropowder, including:

[0096] When the predicted particle size distribution of ultrafine high-purity spherical silicon powder is not within the threshold range of ultrafine high-purity spherical silicon powder particle size distribution, the deviation value between the predicted particle size distribution of ultrafine high-purity spherical silicon powder and the threshold range of ultrafine high-purity spherical silicon powder particle size distribution is extracted.

[0097] The deviation value is compared with the threshold range of particle size distribution of ultrafine high-purity spherical silicon micropowder to obtain the deviation ratio value.

[0098] The real-time data of the preparation process of ultrafine high-purity spherical silicon micropowder was adjusted according to the gradient adjustment based on the initial process real-time data;

[0099] Extract the deviation between the predicted particle size distribution of ultrafine high-purity spherical silicon micropowder after each adjustment and the threshold range of particle size distribution of ultrafine high-purity spherical silicon micropowder;

[0100] The dynamic deviation ratio is obtained by comparing the deviation between the predicted particle size distribution of ultrafine high-purity spherical silicon powder after each adjustment and the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder with the span of the threshold range of particle size distribution of ultrafine high-purity spherical silicon powder.

[0101] Compare the dynamic deviation ratio value with the deviation ratio value;

[0102] When the dynamic deviation ratio exceeds the deviation ratio, the real-time process data adjustment gradient is adjusted to reduce the real-time process data adjustment gradient, preventing excessive fluctuations in the real-time process data adjustment that could lead to poor control stability and imbalance.

[0103] The technical effects of the above solution are as follows: By real-time monitoring of the predicted particle size distribution of ultrafine high-purity spherical silicon micropowder and comparing it with a preset threshold range, it is possible to promptly detect whether the particle size distribution deviates from the target range. Extracting deviation values ​​and performing ratio processing yields a deviation ratio value, providing a precise quantitative basis for subsequent adjustments. Continuously extracting and dynamically comparing the adjusted deviation values ​​ensures that the particle size distribution consistently approaches and remains within the target threshold range, thereby improving the control accuracy of particle size distribution. The initial real-time process data adjustment gradient is set to allow for timely adjustments upon detecting deviations. By dynamically comparing the dynamic deviation ratio value with the actual deviation ratio value, when the dynamic deviation ratio value exceeds the actual deviation ratio value, the real-time process data adjustment gradient is appropriately reduced, which helps prevent process fluctuations caused by over-adjustment. Reducing the adjustment gradient can decrease drastic changes in process parameters, thereby enhancing the stability of the entire preparation process. This technical solution, through real-time data feedback and dynamic adjustment mechanisms, can more efficiently identify and resolve particle size distribution deviation problems. By gradually adjusting and monitoring the adjustment effect, it can be ensured that each adjustment is made in the correct direction, avoiding ineffective or reverse adjustments. The optimized control strategy not only improves control efficiency but also ensures the quality and performance of the final product. Precise control of particle size distribution reduces waste caused by defective products. A stable process also helps reduce equipment failures and downtime, thereby lowering maintenance and production costs.

[0104] In summary, this technical solution significantly improves the control precision of particle size distribution, enhances process stability, optimizes adjustment efficiency and effectiveness, and reduces resource waste and costs. These technical effects are of great significance for improving the production quality and efficiency of ultrafine high-purity spherical silicon micropowder.

[0105] Specifically, when the dynamic deviation ratio exceeds the deviation ratio, the real-time process data adjustment gradient is adjusted, including:

[0106] When the dynamic deviation ratio value exceeds the deviation ratio value, obtain the ratio deviation value between the dynamic deviation ratio value and the deviation ratio value;

[0107] The real-time reaction temperature during the preparation of ultrafine high-purity spherical silicon micropowder is retrieved.

[0108] The real-time reaction temperature in the current preparation process of ultrafine high-purity spherical silicon micropowder is compared with the reference temperature to obtain the temperature deviation value;

[0109] The real-time gas pressure in the reaction chamber during the current preparation process of ultrafine high-purity spherical silicon micropowder was obtained;

[0110] The gas pressure in the real-time reaction chamber during the current preparation process of ultrafine high-purity spherical silicon micropowder is compared with the reference gas pressure to obtain the gas pressure deviation value.

[0111] The temperature deviation value and the gas pressure deviation value are standardized to obtain the standardized temperature deviation value and gas pressure deviation value.

[0112] The gradient adjustment coefficient is obtained by combining the standardized temperature deviation value and gas pressure deviation value with the proportional deviation value between the dynamic deviation ratio value and the deviation ratio value.

[0113] The gradient adjustment coefficient is obtained by the following formula:

[0114]

[0115] Where δ represents the gradient adjustment coefficient; P t This represents the temperature deviation value after standardization; P a B represents the gas pressure deviation value after standardization; D represents the dynamic deviation ratio value and the deviation ratio value;

[0116] The gradient adjustment coefficient is used to adjust the real-time process data adjustment gradient, wherein the adjusted real-time process data adjustment gradient is obtained by the following formula:

[0117]

[0118] Among them, G t G0 represents the adjusted real-time process data adjustment gradient; δ represents the gradient adjustment coefficient; D b The value represents the standard deviation of the dynamic deviation ratio; e represents the preset minimum coefficient to prevent ineffective gradient adjustment when the standard deviation of the dynamic deviation ratio is 0.

[0119] The technical benefits of the above solution are as follows: By monitoring dynamic deviation ratios, temperature deviations, and gas pressure deviations in real time, abnormal fluctuations in the preparation process of ultrafine high-purity spherical silicon micropowder can be detected promptly. Once the dynamic deviation ratio exceeds the set value, a gradient adjustment coefficient is calculated based on the relevant deviation value to adjust the real-time process data adjustment gradient, enabling the process parameters to quickly return to a reasonable range. This reduces product quality instability caused by parameter fluctuations and improves the consistency and stability of product quality.

[0120] This formula comprehensively considers temperature deviation, gas pressure deviation, and the proportional deviation between the dynamic deviation ratio and the deviation ratio to determine the gradient adjustment coefficient. The exponential function exp(-P) t ),exp(-P aThis is used to perform a nonlinear transformation on the standardized temperature deviation and gas pressure deviation values, highlighting the degree of influence of the deviation values ​​on the adjustment coefficient. Taking the maximum of the two values ​​emphasizes the consideration of the deviation factor with the greater impact. This value is then multiplied by the proportional deviation value B and divided by the dynamic deviation proportional value D to obtain the final gradient adjustment coefficient δ. Simultaneously, the gradient adjustment coefficient reflects the degree of adjustment of process parameters when comprehensively considering temperature, gas pressure deviations, and dynamic deviation proportional deviations. A larger δ value indicates a more severe deviation of the current process parameters from the ideal state, requiring a greater adjustment to the real-time process data adjustment gradient.

[0121] By calculating the proportional deviation between the dynamic deviation ratio and the deviation ratio, the degree of deviation between the current process state and the target state can be quantified more accurately, thereby enabling finer adjustments to the real-time process data adjustment gradient. Introducing real-time reaction temperature and reaction chamber gas pressure as adjustment parameters not only considers the influence of particle size distribution but also other process parameters that may affect product quality, enhancing the comprehensiveness and flexibility of the adjustment. Standardizing the temperature and gas pressure deviation values ​​eliminates the influence of dimensions and value ranges between different parameters, making these parameters comparable in gradient adjustment coefficient calculations, thus improving the stability and controllability of process parameters. Using gradient adjustment coefficients to adjust the real-time process data adjustment gradient ensures that each adjustment is based on precise quantitative analysis of the current process state, avoiding the risks of blind and over-adjustment, and improving process stability and controllability. Precise gradient adjustment can reduce raw material waste and energy consumption caused by process parameter fluctuations, thereby improving resource utilization efficiency. Simultaneously, the stability and optimization of process parameters can reduce rework and scrap due to product quality issues, further reducing production costs. By monitoring and adjusting process parameters in real time, the particle size distribution and other performance indicators of ultrafine, high-purity spherical silicon micropowder can be kept within the target range, thereby improving product quality and consistency. This technical solution enhances the system's adaptability and robustness to changes in process conditions by introducing a real-time monitoring and adjustment mechanism for multiple process parameters. Even when process conditions change significantly, the system can maintain the stability and optimization of process parameters through precise gradient adjustment, thus ensuring the stability and consistency of product quality.

[0122] Meanwhile, the above technical solution can reduce the real-time process data adjustment gradient in a timely manner, and maximize the match between the adjusted adjustment gradient and the current process data adjustment reality. This prevents the process adjustment stability from being reduced due to process fluctuations caused by over-adjustment. At the same time, by using the standard deviation of the dynamic deviation ratio during the process adjustment to control the gradient reduction magnitude, the process adjustment efficiency can be controlled to the maximum extent while reducing the fluctuation gradient, preventing the problem of excessively long adjustment cycles caused by excessively low process adjustment efficiency.

[0123] In summary, this technical solution significantly enhances the precision and flexibility of adjustment, improves process stability and controllability, optimizes resource utilization and reduces production costs, improves product quality and consistency, and enhances system adaptability and robustness. These technical effects are of great significance for improving the production quality and efficiency of ultrafine, high-purity spherical silicon micropowder.

[0124] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0125] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A superfine high-purity spherical silicon micro-powder particle size distribution regulation system, characterized in that, The method comprises the following steps: A process parameter acquisition module is configured to acquire real-time data of the superfine high-purity spherical silicon powder preparation process; A process parameter processing module is configured to process the acquired real-time data of the superfine high-purity spherical silicon powder preparation process, and determine the superfine high-purity spherical silicon powder preparation process characteristic data; A particle size distribution regulation module is configured to construct a superfine high-purity spherical silicon powder particle size distribution prediction model, analyze the superfine high-purity spherical silicon powder preparation process characteristic data, predict the superfine high-purity spherical silicon powder particle size distribution, and regulate the superfine high-purity spherical silicon powder particle size distribution in a closed loop; The real-time data of the superfine high-purity spherical silicon powder preparation process is adjusted, including: When the superfine high-purity spherical silicon powder particle size distribution prediction result is not within the superfine high-purity spherical silicon powder particle size distribution threshold range, the deviation value between the superfine high-purity spherical silicon powder particle size distribution prediction result and the superfine high-purity spherical silicon powder particle size distribution threshold range is extracted; The deviation value is processed by ratio with the span of the superfine high-purity spherical silicon powder particle size distribution threshold range to obtain a deviation proportion value; The real-time data of the superfine high-purity spherical silicon powder preparation process is adjusted according to the initial process real-time data adjustment gradient; The deviation value between the superfine high-purity spherical silicon powder particle size distribution prediction result after each adjustment and the superfine high-purity spherical silicon powder particle size distribution threshold range is extracted; The deviation value between the superfine high-purity spherical silicon powder particle size distribution prediction result after each adjustment and the superfine high-purity spherical silicon powder particle size distribution threshold range is processed by ratio with the span of the superfine high-purity spherical silicon powder particle size distribution threshold range to obtain a dynamic deviation proportion value; The dynamic deviation proportion value is compared with the deviation proportion value; When the dynamic deviation proportion value exceeds the deviation proportion value, the process real-time data adjustment gradient is adjusted, including: When the dynamic deviation proportion value exceeds the deviation proportion value, a proportion deviation value between the dynamic deviation proportion value and the deviation proportion value is obtained; The real-time reaction temperature in the current superfine high-purity spherical silicon powder preparation process is called; The real-time reaction temperature in the current superfine high-purity spherical silicon powder preparation process is compared with the reference temperature to obtain a temperature deviation value; The real-time reaction cavity gas pressure in the current superfine high-purity spherical silicon powder preparation process is called; The real-time reaction cavity gas pressure in the current superfine high-purity spherical silicon powder preparation process is compared with the reference gas pressure to obtain a gas pressure deviation value; The temperature deviation value and the gas pressure deviation value are standardized to obtain standardized temperature deviation value and gas pressure deviation value; A gradient adjustment coefficient is obtained by using the standardized temperature deviation value and the gas pressure deviation value in combination with the proportion deviation value between the dynamic deviation proportion value and the deviation proportion value; The gradient adjustment coefficient is obtained by the following formula: ; wherein δ denotes a gradient adjustment coefficient; P t denotes a normalized temperature deviation value; P a denotes a normalized gas pressure deviation value; B denotes a proportional deviation value between a dynamic deviation proportional value and a deviation proportional value; and D denotes a dynamic deviation proportional value. The process real-time data adjustment gradient is adjusted by using the gradient adjustment coefficient, and the adjusted process real-time data adjustment gradient is obtained by the following formula: ; G0+ δG0 t G0+ δG0 b G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δG0 G0+ δ 2. The particle size distribution control system of superfine high-purity spherical silicon micropowder according to claim 1, characterized in that, The real-time data of the superfine high-purity spherical silicon powder preparation process is acquired, including: Real-time monitoring and collection of the particle size of raw materials, melting temperature, melting time, cooling rate, air flow classification speed, air flow pressure, classification machine speed, screen mesh size and vibration frequency in the preparation process of superfine high-purity spherical silicon powder, to determine the real-time data of the superfine high-purity spherical silicon powder preparation process.

3. The ultrafine high purity spherical silicon micropowder particle size distribution control system according to claim 1, wherein, Processing the real-time data of the superfine high-purity spherical silicon powder preparation process, including: Cleaning the real-time data of the superfine high-purity spherical silicon powder preparation process to remove noise that is not useful for controlling the particle size distribution of the superfine high-purity spherical silicon powder. Checking the real-time data of the superfine high-purity spherical silicon powder preparation process to identify repeated values, missing values, and abnormal values, and processing the repeated values, missing values, and abnormal values in the real-time data of the superfine high-purity spherical silicon powder preparation process. Wherein, the repeated values in the real-time data of the superfine high-purity spherical silicon powder preparation process are removed, and it is determined whether the missing values and abnormal values in the real-time data of the superfine high-purity spherical silicon powder preparation process are useful for controlling the particle size distribution of the superfine high-purity spherical silicon powder. If so, the missing values are filled with the average and the abnormal values are replaced with the median. If not, the missing values and abnormal values are directly deleted.

4. The particle size distribution control system of superfine high-purity spherical silicon micropowder according to claim 3, characterized in that, Processing the real-time data of the superfine high-purity spherical silicon powder preparation process also includes: Normalizing the real-time data of the superfine high-purity spherical silicon powder preparation process to convert it to a unified data format, remove dimensional differences in the real-time data of the superfine high-purity spherical silicon powder preparation process, and determine standardized real-time data of the superfine high-purity spherical silicon powder preparation process. Feature extraction from the real-time data of the superfine high-purity spherical silicon powder preparation process to extract features useful for controlling the particle size distribution of the superfine high-purity spherical silicon powder from the real-time data of the superfine high-purity spherical silicon powder preparation process, and determine the superfine high-purity spherical silicon powder preparation process feature data.

5. The ultrafine high purity spherical silicon micropowder particle size distribution control system of claim 1, wherein, Building a superfine high-purity spherical silicon powder particle size distribution prediction model, including: Collecting superfine high-purity spherical silicon powder preparation process historical data based on the superfine high-purity spherical silicon powder particle size distribution control requirements, including the superfine high-purity spherical silicon powder particle size distribution situation; Dividing the collected superfine high-purity spherical silicon powder preparation process historical data into a training set and a test set; Based on deep learning technology, using the training set to train the deep learning model, so that the deep learning model can learn the relationship between the superfine high-purity spherical silicon powder particle size distribution and the superfine high-purity spherical silicon powder preparation process, and automatically learn the superfine high-purity spherical silicon powder particle size distribution prediction behavior, to determine the deep learning-based superfine high-purity spherical silicon powder particle size distribution prediction model; Using the test set to test the performance of the deep learning-based superfine high-purity spherical silicon powder particle size distribution prediction model to evaluate whether the deep learning-based superfine high-purity spherical silicon powder particle size distribution prediction model can achieve the effect of predicting the superfine high-purity spherical silicon powder particle size distribution. When the deep learning-based ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model cannot achieve the effect of predicting the particle size distribution of the ultra-fine high-purity spherical silicon micro-powder, the parameters of the deep learning-based ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model are adjusted, and the deep learning-based ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is continuously optimized until the deep learning-based ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model can achieve the effect of predicting the particle size distribution of the ultra-fine high-purity spherical silicon micro-powder, and the best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined.

6. The particle size distribution control system of superfine high-purity spherical silicon micropowder according to claim 5, characterized in that, The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model. The best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is obtained, and the best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is deployed in the actual ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction environment. The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is input into the best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is analyzed based on the best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is analyzed based on the best ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model.

7. The ultrafine high-purity spherical silicon micropowder particle size distribution regulation system according to claim 6, characterized in that, The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model. The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model. The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model. The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model.

8. A method for controlling the particle size distribution of ultra-fine high-purity spherical silicon powder, based on a system for controlling the particle size distribution of ultra-fine high-purity spherical silicon powder according to claim 7, characterized in that, The ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model, and the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model is determined based on the analysis of the ultra-fine high-purity spherical silicon micro-powder particle size distribution prediction model. ​ ​ ​

Images