Parameter optimal configuration method, device, system and equipment for raw material of polysilicon production

Abstract

The application provides a parameter optimization configuration method, device, system and equipment for a polysilicon production raw material, wherein the method comprises the following steps: obtaining real-time reaction data in current polysilicon production; obtaining preselected raw material data according to the real-time reaction data and a pre-constructed product model; obtaining at least one raw material parameter configuration information according to the preselected raw material data and a pre-constructed parameter optimization prediction model; and determining and outputting one of the at least one raw material parameter configuration information as parameter optimization configuration information according to current production process information, equipment state information and power supply demand information. The scheme of the application reduces the actual problems of unstable and insufficient equipment reaction caused by material reaction in raw materials and large fluctuation of catalyst efficiency, improves the economy and safety of internal full reaction of the equipment, and effectively relieves the pressure caused by the rise of production cost.

Description

Technical Field

[0001] This application relates to the field of polysilicon production control technology, and in particular to a method, apparatus, system and equipment for the optimal configuration of parameters for polysilicon production raw materials. Background Technology

[0002] The first and crucial step in the production of electronic-grade polysilicon is the production of trichlorosilane. However, the raw materials and catalysts for this step are all sourced externally. Externally sourced solid raw materials, such as silicon powder and cuprous chloride catalyst, suffer from issues related to different suppliers, origins, mesh size variations, and differences in purity and moisture content. The purity of externally sourced liquid silicon tetrachloride also significantly impacts production, as does the purity and moisture content of hydrogen chloride gas, which affects the conversion rate. To reduce costs and improve economic efficiency, traditional enterprises have adopted a long-term strategy of selecting relatively fixed suppliers and parameters based on historical experience to cope with the external market. However, in today's highly competitive environment, using raw materials from different manufacturers and with different parameters to further improve the conversion of chlorine in the cold hydrogenation workshop and reduce its energy consumption will make a significant contribution to the ultimate goal of cost reduction in polysilicon production. Therefore, the selection of optimal parameters for raw material proportioning urgently requires a raw material decision-making and proportioning management solution that combines a full-cost proportioning model with a systematic implementation system. This solution is of great significance for promoting energy conservation and consumption reduction in polysilicon production. Summary of the Invention

[0003] The technical objective of this application is to provide a method, apparatus, system, and equipment for optimally configuring parameters of raw materials for polysilicon production, in order to solve the problem that the current raw material ratios of polysilicon manufacturers cannot meet the requirements for energy conservation and consumption reduction.

[0004] To address the aforementioned technical problems, embodiments of this application provide a method for optimally configuring parameters of raw materials for polysilicon production, comprising:

[0005] Obtain real-time reaction data in the current polysilicon production process;

[0006] Based on the real-time reaction data and the pre-built product model, pre-selected raw material data is obtained, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective.

[0007] Based on the pre-selected raw material data and the pre-built parameter optimization prediction model, at least one raw material parameter configuration information is obtained, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status, and power generation load;

[0008] Based on the current production process information, equipment status information, and power supply demand information, determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information.

[0009] Specifically, as described above, constructing the product model includes:

[0010] The historical raw material data, historical reaction data, and historical product data are acquired. The historical raw material data includes raw material parameters and raw material supply information. The raw material parameters include at least one of type and quality. The raw material supply information includes at least one of supplier, place of origin, batch, and supply price.

[0011] Based on the model objective and preset constraints, an initial neural network model is constructed, wherein the preset constraints include at least one of the following: the extreme value of the ratio of each raw material, the maximum amount of each raw material used, and the minimum quality limit value of each raw material.

[0012] The initial neural network model is trained and tested based on the historical raw material data, the historical reaction data, and the historical product data to obtain the product model.

[0013] Specifically, in the method described above, the reaction data and / or the historical reaction data include at least one of the following: reaction temperature, reaction rate, calorific value, moisture content, volatile matter, product purity, energy utilization rate, and deimpurification capability.

[0014] Specifically, the method described above, for constructing the parameter optimization prediction model, includes:

[0015] Based on the degree of influence of each raw material parameter on the production process, the characteristic values ​​of each raw material parameter are determined;

[0016] Based on the correlation between the raw material parameters and the feature values, an initial prediction model is constructed using a support vector machine.

[0017] Based on preset ratios of various raw materials and historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

[0018] Furthermore, in the method described above, after determining the characteristic values ​​of each of the raw material parameters, the method further includes:

[0019] Based on the characteristic value, the raw material parameter whose characteristic value is greater than a preset value is determined to be a key parameter;

[0020] Based on the preset ratios of various key parameters and the historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

[0021] Optionally, the method described above further includes:

[0022] Receive raw material supply information from the digital delivery platform;

[0023] Obtain the raw material parameters of the raw materials after they have passed the quality inspection process;

[0024] The raw material parameters and raw material supply information corresponding to the same raw material are bound together to obtain raw material data.

[0025] Another embodiment of this application also provides a control device, including:

[0026] The first processing module is used to acquire real-time reaction data in the current polysilicon production process.

[0027] The second processing module is used to obtain pre-selected raw material data based on the real-time reaction data and the pre-built product model, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective.

[0028] The third processing module is used to obtain at least one raw material parameter configuration information based on the pre-selected raw material data and the pre-built parameter optimization prediction model, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status and power generation load.

[0029] The fourth processing module is used to determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information based on the current production process information, equipment status information and power supply demand information.

[0030] Another embodiment of this application provides a control system, including:

[0031] The equipment layer is used to acquire information collected by various devices in polysilicon production.

[0032] The edge layer is used to provide connectivity-as-a-service capabilities and real-time computing processing for industrial IoT.

[0033] The platform layer is used to integrate basic capabilities including Platform as a Service, data and services, and industrial IoT platforms.

[0034] The application layer is used to implement the steps of the method for optimally configuring parameters of polysilicon production raw materials as described above.

[0035] Another embodiment of this application provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described above.

[0036] Another embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described above.

[0037] Compared with the prior art, the preferred configuration method, apparatus, system and equipment for polysilicon production raw materials provided in this application have at least the following beneficial effects:

[0038] This application uses a pre-built product model and parameter optimization prediction model, and calculates based on real-time reaction data, to determine the optimal parameter configuration information that is suitable for current production. This allows production to be carried out based on the optimal parameter configuration information. Through digital lean management, it reduces practical problems such as unstable and incomplete equipment reactions caused by large fluctuations in material reactions and catalyst efficiency in raw materials. It improves the economy and safety of full reaction within the equipment and effectively alleviates the pressure caused by rising production costs. Attached Figure Description

[0039] Figure 1 This is one of the flowcharts illustrating the method for optimally configuring parameters of raw materials for polysilicon production in this application;

[0040] Figure 2 The second flowchart illustrates the method for optimally configuring parameters of raw materials for polysilicon production in this application.

[0041] Figure 3 The third flowchart illustrates the method for optimizing the parameters of raw materials for polysilicon production in this application.

[0042] Figure 4 The fourth flowchart illustrates the method for optimizing the parameters of raw materials for polysilicon production in this application.

[0043] Figure 5 The fifth flowchart illustrates the method for optimally configuring parameters of raw materials for polysilicon production in this application.

[0044] Figure 6 This is a schematic diagram of the control device in this application. Detailed Implementation

[0045] To make the technical problems, technical solutions, and advantages of this application clearer, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments. In the following description, specific details such as particular configurations and components are provided merely to aid in a comprehensive understanding of the embodiments of this application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Furthermore, for clarity and brevity, descriptions of known functions and structures have been omitted.

[0046] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0047] In the various embodiments of this application, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0048] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0049] In the embodiments provided in this application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0050] It should be noted that this application uses the production of trichlorosilane in polysilicon manufacturing as an example.

[0051] See Figure 1 One embodiment of this application provides a method for optimally configuring parameters of raw materials for polycrystalline silicon production, including:

[0052] Step S101: Obtain real-time reaction data during the current polysilicon production process;

[0053] Step S102: Based on the real-time reaction data and the pre-built product model, pre-selected raw material data is obtained, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective.

[0054] Step S103: Based on the pre-selected raw material data and the pre-constructed parameter optimization prediction model, at least one raw material parameter configuration information is obtained, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status, and power generation load.

[0055] Step S104: Based on the current production process information, equipment status information, and power supply demand information, determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information.

[0056] In this embodiment, when determining the optimal configuration of raw material parameters, real-time reaction data from the current polysilicon production is used as an indicator parameter and input into a pre-built product model for calculation to obtain pre-selected raw material data under the current conditions. Since the product model takes at least one of product purity, product quantity, reaction energy consumption, and cost as its model objective, it can ensure that the most economical product can be obtained when producing according to the raw materials corresponding to the pre-selected raw material data. After obtaining the pre-selected raw material data, in order to further improve the economy of raw material parameter configuration, it is also substituted into a pre-built parameter optimization prediction model to predict the pre-selected raw material data. Based on at least one of the production process, equipment status, and power generation load, multiple raw material parameter configuration information is obtained. Furthermore, based on the user's current production process information, equipment status information, and power supply demand information, the most economical raw material parameter configuration information is determined from the multiple raw material parameter configuration information as the optimal parameter configuration information, and the optimal parameter configuration information is output so that the user can produce according to the optimal parameter configuration information.

[0057] In summary, this application uses a pre-constructed product model and parameter optimization prediction model, and calculates based on real-time reaction data, to determine the optimal parameter configuration information that is suitable for current production. This allows for production based on the optimal parameter configuration information. Through digital lean management, it reduces practical problems such as unstable and incomplete equipment reactions caused by large fluctuations in material reactions and catalyst efficiency in raw materials. This improves the economy and safety of full reactions within the equipment and effectively alleviates the pressure caused by rising production costs.

[0058] See Figure 2 Specifically, as described above, constructing the product model includes:

[0059] Step S201: Obtain the historical raw material data, the historical reaction data, and the historical product data, wherein the historical raw material data includes: raw material parameters and raw material supply information, the raw material parameters include at least one of type and quality, and the raw material supply information includes at least one of supplier, place of origin, batch and supply price;

[0060] Step S202: Construct an initial neural network model based on the model objective and preset constraints, wherein the preset constraints include at least one of the following: extreme values ​​of the proportion of each raw material, maximum usage of each raw material, and minimum quality limit of each raw material.

[0061] Step S203: Train and test the initial neural network model based on the historical raw material data, the historical reaction data, and the historical product data to obtain the product model.

[0062] In this embodiment, when pre-constructing the product model, historical data related to the production process is first acquired, including: historical raw material data, historical reaction data, and historical product data. The historical raw material data includes: raw material parameters and raw material supply information. The raw material parameters include at least one of: type (the specific name of the raw material or catalyst) and quality (including at least one of purity, moisture content, and mesh size). The raw material supply information includes at least one of: supplier, origin, batch number, and supply price. The reaction data and / or historical reaction data include at least one of: reaction temperature, reaction rate, calorific value, moisture content, volatile matter, product purity, energy utilization rate, and impurity removal capability. The historical product data includes at least one of: product quantity, product purity, product reaction energy consumption, and product production cost.

[0063] Based on the relationship between raw material data, reaction data, and product data, and after determining the model objective and preset constraints, an initial neural network model is constructed. The model objective includes at least one of the following: product purity, product quantity, reaction energy consumption, and cost. The preset constraints include at least one of the following: extreme values ​​of the proportions of each raw material, maximum usage of each raw material, and minimum quality limits for each raw material.

[0064] Furthermore, historical data is divided into training and testing sets according to a preset ratio. The initial neural network model is first trained based on the data in the training set. After training, the trained initial neural network model is tested and optimized again based on the data in the testing set to ensure that the final product model meets the requirements of economy and security.

[0065] It should be noted that the raw material parameters and raw material supply information for each raw material in the raw material data are linked.

[0066] See Figure 3 Specifically, the method described above, for constructing the parameter optimization prediction model, includes:

[0067] Step S301: Determine the characteristic values ​​of each raw material parameter based on the degree of influence of each raw material parameter on the production process;

[0068] Step S302: Based on the correlation between the raw material parameters and the feature values, construct an initial prediction model using a support vector machine.

[0069] Step S303: Based on the preset ratios of various raw materials and historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

[0070] In this embodiment, when constructing the parameter optimization prediction model, the feature values ​​corresponding to each raw material parameter are determined based on the degree of influence of each raw material parameter on the production process. The greater the influence of a raw material parameter on the production process, the larger the determined feature value. Then, based on the correlation between the raw material parameters and the feature values, an initial prediction model based on a support vector machine is constructed. Furthermore, based on preset ratios of various raw materials and historical raw material data, the initial prediction model is trained and optimized to obtain the desired parameter optimization prediction model.

[0071] It should be noted that the steps for training and optimizing the initial prediction model are similar to the steps for training and testing the product model described above. That is, historical data is divided into training and testing sets. Training is performed using data from the training set, and testing and optimization are performed using data from the testing set to obtain the desired parameter-optimized prediction model. Furthermore, when optimizing the parameter-optimized prediction model, optimization is based on a population evolution algorithm.

[0072] See Figure 4 Furthermore, in the method described above, after determining the characteristic values ​​of each of the raw material parameters, the method further includes:

[0073] Step S401: Determine the raw material parameter whose feature value is greater than a preset value as a key parameter based on the feature value;

[0074] Step S402: Based on the preset ratios of various key parameters and the historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

[0075] In this embodiment, after determining the feature values ​​of each raw material parameter, the raw material parameters are screened based on the feature values. The raw material parameters with feature values ​​greater than preset values ​​are identified as key parameters. Then, during training optimization, training optimization is mainly based on the ratio of multiple key parameters, which helps to reduce the number of parameters designed in the calculation process and thus improves the calculation efficiency.

[0076] See Figure 5 Optionally, the method described above further includes:

[0077] Step S501: Receive raw material supply information provided by the digital delivery platform;

[0078] Step S502: Obtain the raw material parameters after the raw material has passed the quality inspection process;

[0079] Step S503: Bind the raw material parameters and the raw material supply information corresponding to the same raw material to obtain raw material data.

[0080] In this embodiment, when acquiring raw material data, the raw material supply information described above, provided by the digital delivery platform, is received. To ensure the accuracy of the raw material parameter information, a quality inspection process is performed to determine the required raw material parameters. Preferably, this quality inspection process is conducted through a Laboratory Information Management System (LIMS) integrated into the Industrial Internet platform. After acquiring the raw material parameters, the raw material parameters and supply information corresponding to the same raw material are bound together to obtain the aforementioned raw material data.

[0081] The digital delivery platform includes, but is not limited to, warehouse management systems (WMS) and procurement management systems.

[0082] See Figure 6 Another embodiment of this application also provides a control device, including:

[0083] The first processing module 601 is used to acquire real-time reaction data in the current polysilicon production process.

[0084] The second processing module 602 is used to obtain pre-selected raw material data based on the real-time reaction data and the pre-built product model, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective.

[0085] The third processing module 603 is used to obtain at least one raw material parameter configuration information based on the pre-selected raw material data and the pre-constructed parameter optimization prediction model, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status and power generation load.

[0086] The fourth processing module 604 is used to determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information based on the current production process information, equipment status information and power supply demand information.

[0087] Specifically, the device described above includes:

[0088] The fifth processing module is used to acquire the historical raw material data, the historical reaction data, and the historical product data. The historical raw material data includes raw material parameters and raw material supply information. The raw material parameters include at least one of type and quality. The raw material supply information includes at least one of supplier, place of origin, batch, and supply price.

[0089] The sixth processing module is used to construct an initial neural network model based on the model objective and preset constraints, wherein the preset constraints include at least one of the following: extreme values ​​of the proportion of each raw material, maximum usage of each raw material, and minimum quality limit of each raw material.

[0090] The seventh processing module is used to train and test the initial neural network model based on the historical raw material data, the historical reaction data, and the historical product data to obtain the product model.

[0091] Specifically, in the apparatus described above, the reaction data and / or the historical reaction data include at least one of the following: reaction temperature, reaction rate, calorific value, moisture content, volatile matter, product purity, energy utilization rate, and deimpurification capability.

[0092] Specifically, the device described above includes:

[0093] The eighth processing module is used to determine the characteristic values ​​of each raw material parameter based on the degree of influence of each raw material parameter on the production process;

[0094] The ninth processing module is used to construct an initial prediction model based on a support vector machine according to the correlation between the raw material parameters and the feature values.

[0095] The tenth processing module is used to train and optimize the initial prediction model based on preset ratios of various raw materials and historical raw material data to obtain the parameter-optimized prediction model.

[0096] Furthermore, the apparatus described above also includes:

[0097] The eleventh processing module is used to determine, based on the characteristic value, that the raw material parameter whose characteristic value is greater than a preset value is a key parameter;

[0098] The twelfth processing module is used to train and optimize the initial prediction model based on the preset ratio of multiple key parameters and the historical raw material data, so as to obtain the parameter-optimized prediction model.

[0099] Optionally, the apparatus described above further includes:

[0100] The thirteenth processing module is used to receive raw material supply information provided by the digital delivery platform;

[0101] The fourteenth processing module is used to obtain the raw material parameters of the raw materials after they have passed the quality inspection process;

[0102] The fifteenth processing module is used to bind the raw material parameters and the raw material supply information corresponding to the same raw material to obtain raw material data.

[0103] The apparatus embodiment of this application is an apparatus corresponding to the embodiment of the above-described method for optimal configuration of parameters of polysilicon production raw materials. All implementation means in the above-described method embodiment are applicable to the apparatus embodiment and can achieve the same technical effect.

[0104] Based on the Industrial Internet platform, the overall deployment of the digital intelligent polysilicon production raw material optimal parameter analysis and configuration solution is divided into four layers: equipment layer, edge layer, platform layer, and application layer.

[0105] Another embodiment of this application provides a control system based on an industrial internet platform. This control system deploys a method for optimizing the configuration of parameters for polysilicon production raw materials across four layers, including:

[0106] The equipment layer is used to acquire information collected by various devices in polysilicon production. For example, in the production of trichlorosilane, it connects to intelligent devices, integrated systems, and information systems in the cold hydrogenation process.

[0107] The edge layer is used to provide connectivity-as-a-service capabilities and real-time computing processing for industrial IoT. It can connect to edge computing devices such as data acquisition gateways and protocol conversion gateways deployed with hyperconverged server resources.

[0108] The platform layer integrates foundational capabilities including Platform as a Service (PASS), Data as a Service (DAAS), and Industrial Internet of Things (IIOT) platforms. The PASS layer provides services such as deployment and elastic resource scaling. The IIIOT platform aggregates data and provides basic data support to the DAAS layer. The DAAS layer centrally manages data resources through a data warehouse and provides interfaces for Artificial Intelligence (AI) models.

[0109] The application layer is used to implement the steps of the parameter optimization configuration method for polysilicon production raw materials as described above. The application layer focuses on the entire scenario and life cycle of the polysilicon production industry and provides several independently deployable industrial applications, such as the cold hydrogenation industrial production material analysis system, the reduction industrial ratio optimization application, and the distillation process analysis system.

[0110] Another embodiment of this application provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described above.

[0111] Another embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described above.

[0112] Furthermore, reference numerals and / or letters may be repeated in different examples within this application. Such repetition is for the purpose of simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or settings discussed.

[0113] It should also 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.

[0114] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for optimally configuring parameters of raw materials for polycrystalline silicon production, characterized in that, include: Obtain real-time reaction data in the current polysilicon production process; Based on the real-time reaction data and the pre-built product model, pre-selected raw material data is obtained, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective. Based on the pre-selected raw material data and the pre-built parameter optimization prediction model, at least one raw material parameter configuration information is obtained, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status, and power generation load; Based on the current production process information, equipment status information, and power supply demand information, determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information; Constructing the parameter optimization prediction model includes: Based on the degree of influence of each raw material parameter on the production process, the characteristic values ​​of each raw material parameter are determined; Based on the correlation between the raw material parameters and the feature values, an initial prediction model is constructed using a support vector machine. Based on preset ratios of various raw materials and historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

2. The method according to claim 1, characterized in that, Constructing the product model includes: The historical raw material data, historical reaction data, and historical product data are acquired. The historical raw material data includes raw material parameters and raw material supply information. The raw material parameters include at least one of type and quality. The raw material supply information includes at least one of supplier, place of origin, batch, and supply price. Based on the model objective and preset constraints, an initial neural network model is constructed, wherein the preset constraints include at least one of the following: the extreme value of the ratio of each raw material, the maximum amount of each raw material used, and the minimum quality limit value of each raw material. The initial neural network model is trained and tested based on the historical raw material data, the historical reaction data, and the historical product data to obtain the product model.

3. The method according to claim 1 or 2, characterized in that, The reaction data and / or the historical reaction data include at least one of the following: reaction temperature, reaction rate, calorific value, moisture content, volatile matter, product purity, energy utilization rate, and deimpurification capability.

4. The method according to claim 1, characterized in that, After determining the characteristic values ​​of each of the raw material parameters, the method further includes: Based on the characteristic value, the raw material parameter whose characteristic value is greater than a preset value is determined to be a key parameter; Based on the preset ratios of various key parameters and the historical raw material data, the initial prediction model is trained and optimized to obtain the parameter-optimized prediction model.

5. The method according to claim 2, characterized in that, Also includes: Receive raw material supply information from the digital delivery platform; Obtain the raw material parameters of the raw materials after they have passed the quality inspection process; The raw material parameters and raw material supply information corresponding to the same raw material are bound together to obtain raw material data.

6. A control device, characterized in that, include: The first processing module is used to acquire real-time reaction data in the current polysilicon production process. The second processing module is used to obtain pre-selected raw material data based on the real-time reaction data and the pre-built product model, wherein the product model is constructed based on historical raw material data and historical reaction data, and the product model takes at least one of product purity, product quantity, reaction energy consumption and cost as the model objective. The third processing module is used to obtain at least one raw material parameter configuration information based on the pre-selected raw material data and the pre-built parameter optimization prediction model, wherein the parameter optimization prediction model is constructed based on the correlation between each raw material parameter, and the at least one raw material parameter configuration information is distinguished according to at least one of the production process, equipment status and power generation load. The fourth processing module is used to determine and output one of the at least one raw material parameter configuration information as the preferred parameter configuration information based on the current production process information, equipment status information and power supply demand information; It also includes: an eighth processing module, used to determine the characteristic values ​​of each raw material parameter based on the degree of influence of each raw material parameter on the production process; The ninth processing module is used to construct an initial prediction model based on a support vector machine according to the correlation between the raw material parameters and the feature values. The tenth processing module is used to train and optimize the initial prediction model based on preset ratios of various raw materials and historical raw material data to obtain the parameter-optimized prediction model.

7. A control system, characterized in that, include: The equipment layer is used to acquire information collected by various devices in polysilicon production. The edge layer is used to provide connectivity-as-a-service capabilities and real-time computing processing for industrial IoT. The platform layer is used to integrate basic capabilities including Platform as a Service, data and services, and industrial IoT platforms. An application layer for implementing the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described in any one of claims 1 to 5.

8. An electronic device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described in any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the method for optimally configuring parameters of polycrystalline silicon production raw materials as described in any one of claims 1 to 5.

Images