Method, system and equipment for optimizing batching parameters in industrial silicon smelting process

Abstract

The invention discloses a method, a system and equipment for optimizing batching parameters in an industrial silicon smelting process. The method comprises the following steps: firstly, acquiring process parameters in an industrial silicon smelting process, standardizing the process parameters, then converting the process parameters into weight factors influencing a smelting result, and mapping a relation between the smelting result and the weight factors into a parameter equation based on a preset multiple linear regression model, obtaining a parameter set containing a weight factor and a linear regression coefficient; then, based on the parameter set, the initial raw material parameters and the reaction process parameters, an equivalent circuit model is constructed and simulated to generate smelting result data under different parameter combinations, and therefore parameter adjustment variables are obtained; according to the numerical range of the parameter adjustment variables and the initial raw material parameters, an optimal batching parameter combination is obtained; and finally, performing optimization treatment on batching parameters in the industrial silicon smelting process. According to the method, accurate control and rapid optimization adjustment of batching parameters can be stably achieved in the industrial silicon smelting process.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial silicon smelting, and in particular relates to a method and system for optimizing batching parameters in an industrial silicon smelting process, and industrial silicon production equipment. Background Art

[0002] In the process of industrial silicon smelting, optimizing the batching parameters is the key to ensuring smelting efficiency and finished product quality. However, current technologies mainly rely on empirical rules and test methods to select and optimize these parameters, which often require a lot of experiments and repeated adjustments. This is not only time-consuming and labor-intensive, but also difficult to stably achieve the best smelting effect. In addition, the impact of factory environment, climatic conditions and equipment type on the smelting process is often ignored, or simple adjustments are made based on experience, resulting in instability in the production process and fluctuations in product quality. For example, in a high humidity environment, the materials in the reactor may be difficult to fully react, and in extreme temperature conditions, the operating efficiency of the equipment will also be affected, thereby affecting the control accuracy of the batching reaction.

[0003] Although some attempts have been made in existing technologies, such as the introduction of advanced sensors and automated control systems to monitor and adjust batching parameters in real time, these methods often fail to fully consider the impact of complex factory environments and climate change, resulting in unstable smelting results. At the same time, existing empirical methods and experimental methods lack systematicity and scientificity, and are difficult to cope with the complex and changing environment and process requirements in the process of industrial silicon smelting, resulting in inefficient optimization processes and difficulty in achieving ideal smelting results. Summary of the invention

[0004] The technical problem to be solved by the present invention is to address the above-mentioned deficiencies in the prior art and to propose a method, system and industrial silicon production equipment for optimizing the batching parameters in the industrial silicon smelting process. The method optimizes the batching parameters, thereby being able to stably achieve precise control and rapid optimization and adjustment of the batching parameters in the industrial silicon smelting process, and more effectively cope with the complex and changeable environment and process requirements in industrial silicon smelting.

[0005] In a first aspect, the present invention provides a method for optimizing batching parameters in an industrial silicon smelting process, the method comprising the following steps:

[0006] Step S1: obtaining process parameters in an industrial silicon smelting process; and standardizing the process parameters;

[0007] The process parameters include climatic condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters;

[0008] Step S2: converting the standardized process parameters into weight factors that affect the smelting results, and mapping the relationship between the smelting results and the weight factors into a parameter equation based on a preset multivariate linear regression model, thereby obtaining a parameter set including weight factors and linear regression coefficients;

[0009] Step S3: constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters; and simulating the industrial silicon smelting process based on the equivalent circuit model to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables;

[0010] Step S4: according to the numerical range of the parameter adjustment variable and the initial raw material parameters, an optimal batching parameter combination is obtained;

[0011] Step S5: optimizing the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

[0012] In a second aspect, the present invention provides a system for optimizing batching parameters in an industrial silicon smelting process, the system comprising:

[0013] An acquisition module is used to acquire process parameters in the industrial silicon smelting process; the process parameters include climate condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters;

[0014] A standardization module, connected to the acquisition module, for performing standardization processing on the process parameters;

[0015] A first processing module, connected to the standardization module, is used to convert the standardized process parameters into weight factors that affect the smelting results, and based on a preset multivariate linear regression model, map the relationship between the smelting results and the weight factors into a parameter equation, thereby obtaining a parameter set including weight factors and linear regression coefficients;

[0016] A construction module, connected to the first processing module, for constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters;

[0017] A generating module, connected to the building module, for simulating the industrial silicon smelting process based on the equivalent circuit model to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables;

[0018] An optimization module, connected to the generation module, for adjusting the numerical range of the variables and the initial raw material parameters according to the parameters to obtain an optimal batching parameter combination;

[0019] The second processing module is connected to the optimization module and is used to optimize the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

[0020] In a third aspect, the present invention provides an industrial silicon production equipment, characterized in that the industrial silicon production equipment comprises the system for optimizing batching parameters in the industrial silicon smelting process described in the second aspect.

[0021] The present invention comprehensively considers multiple factors such as factory environment, climate conditions and equipment type, and uses mathematical models and intelligent algorithms to optimize the batching parameters, so as to stably achieve accurate control and rapid optimization and adjustment of the batching parameters during the industrial silicon smelting process, thereby more effectively coping with the complex and changeable environment and process requirements in industrial silicon smelting. The specific beneficial effects are as follows:

[0022] 1. Improve production efficiency

[0023] After optimizing the batching parameters, the present invention can accurately control the batching parameters, reduce downtime and material waste in the smelting process, and significantly improve the operating efficiency of the industrial silicon smelting process.

[0024] 2. Ensure product quality

[0025] By optimizing the batching parameters, the present invention can improve the consistency and reliability of products, reduce the product defect rate, and ensure that the final product meets industry standards and customer needs.

[0026] 3. Reduce production costs

[0027] The present invention flexibly uses different raw materials and ingredient combinations, which can reduce the use of high-cost raw materials, thereby significantly reducing overall production costs and improving corporate profitability.

[0028] 4. Quickly adapt to market changes

[0029] The present invention can adjust the batching parameters in real time and respond quickly to market demands or environmental changes, thereby enhancing the ability of enterprises to cope with market fluctuations.

[0030] 5. Reduce environmental impact

[0031] The present invention optimizes the raw materials and energy consumption required for the smelting process, reduces the emission of harmful substances, meets the goal of sustainable development, and enhances the environmentally friendly image of the enterprise.

[0032] 6. Data-driven decision support

[0033] The present invention is based on the optimization of multiple linear regression model and intelligent algorithm, which can provide a scientific basis for decision-making, thereby reducing human errors and improving management level.

[0034] 7. Strengthen technological innovation capabilities

[0035] The present invention introduces advanced mathematical models and intelligent analysis technology, which promotes the technological progress of enterprises in the field of silicon smelting and enhances the market competitiveness of enterprises.

[0036] 8. Achieve systematic management

[0037] The present invention comprehensively considers various parameters and establishes a model, which can systematically manage and optimize the entire smelting process and promote the coordinated work of various links.

[0038] 9. Promote efficient use of resources

[0039] By optimizing the batching parameters, the present invention can make full use of various resources, minimize resource waste, and achieve a win-win situation in economic and environmental benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a schematic diagram of a method for optimizing batching parameters in an industrial silicon smelting process in an embodiment of the present invention;

[0041] Figure 2 is a schematic diagram of an equivalent circuit model in an embodiment of the present invention;

[0042] Figure 3 This is a schematic diagram of a device for optimizing batching parameters in an industrial silicon smelting process in an embodiment of the present invention;

[0043] Among them, the figure numbers are: 10, acquisition module, 20, standardization module, 30, first processing module, 40, construction module, 50, generation module, 60, optimization module, 70, second processing module. DETAILED DESCRIPTION

[0044] In order to enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0045] It should be understood that the specific embodiments and drawings described herein are only used to explain the present invention rather than to limit the present invention.

[0046] It can be understood that, in the absence of conflict, the various embodiments of the present invention and the various features in the embodiments can be combined with each other.

[0047] It can be understood that, for the convenience of description, the drawings of the present invention only show the parts related to the present invention, while the parts irrelevant to the present invention are not shown in the drawings.

[0048] It can be understood that each unit and module involved in the embodiments of the present invention may correspond to only one physical structure, or may be composed of multiple physical structures, or multiple units and modules may be integrated into one physical structure.

[0049] It can be understood that, in the absence of conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the drawings.

[0050] It is understood that the flowcharts and block diagrams of the present invention illustrate the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to the embodiments of the present invention. Each box in the flowchart or block diagram may represent a unit, module, program segment, or code, which contains executable instructions for implementing the specified functions. Moreover, each box or combination of boxes in the block diagram and flowchart may be implemented by a hardware-based system that implements the specified functions, or by a combination of hardware and computer instructions.

[0051] It can be understood that the units and modules involved in the embodiments of the present invention can be implemented by software or hardware. For example, the units and modules can be located in a processor.

[0052] Embodiment 1:

[0053] like Figure 1 As shown, this embodiment provides a method for optimizing batching parameters in an industrial silicon smelting process, the method comprising the following steps:

[0054] Step S1: Acquire process parameters in the process of industrial silicon smelting; and standardize the process parameters; the process parameters include climate condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters.

[0055] For example, the initial raw material parameters include the initial purity of the silica ore and the corresponding ingredient composition and proportion.

[0056] In the process of industrial silicon smelting, reaction process parameters refer to the corresponding key variables and conditions in the smelting and purification reaction of silicon. These parameters directly affect the efficiency of the reaction and the quality of the final product. Specifically, reaction process parameters include but are not limited to the following:

[0057] (1) Reaction temperature: The operating temperature required during the smelting process. The temperature affects the reaction rate and silicon output rate.

[0058] (2) Reaction time: The length of time required for silicon smelting and purification reflects the duration of the reaction process. Different reaction times will result in different product quality and output.

[0059] (3) Gas phase composition: gases involved in the reaction (such as hydrogen, carbon monoxide, etc.) and their concentrations. These gases may participate in the reaction or affect the temperature and material transfer during the smelting process.

[0060] (4) Solid raw material composition: The chemical composition and particle size distribution of the solid raw materials involved in the reaction (such as quartz sand, reducing agent, etc.) affect the reaction effect and the final purity of silicon.

[0061] (5) Charge ratio: The ratio between different raw materials (such as quartz sand, reducing agent, etc.) affects the thermal balance of the reaction and the properties of the products.

[0062] (6) Gas flow rate: The gas flow rate involved in the reaction affects the gas transmission efficiency and local atmosphere during the reaction.

[0063] (7) Pressure: The environmental pressure during the reaction may affect the reaction rate and product yield.

[0064] (8) Cooling rate: The cooling rate of metallic silicon after smelting affects the crystal structure and physical properties of the product.

[0065] Specifically, the standardization processing includes data cleaning of the climate condition parameters, the factory environment parameters, the equipment type parameters, the initial raw material parameters and the reaction process parameters, respectively, to obtain standardized climate condition parameters, standardized factory environment parameters, standardized equipment type parameters, standardized initial raw material parameters and standardized reaction process parameters.

[0066] Standardized parameters refer to the result of converting original parameters of different units or dimensions into a unified scale (usually a range of 0 to 1, or a distribution with a mean of 0 and a variance of 1) through a certain method. This process usually includes the following steps:

[0067] (1) Data cleaning: Process missing values, outliers, and duplicate data to ensure data integrity and accuracy.

[0068] (2) Standardization processing method: This embodiment adopts Min-Max standardization or Z-score standardization. Min-Max standardization is to convert the data into the range of [0, 1], and Z-score standardization is to convert the data into a distribution with a mean of 0 and a standard deviation of 1.

[0069] When standardizing parameters, the standardized parameters need to meet the following conditions:

[0070] (1) Consistency: The standardized parameters should be comparable and can be used to compare different data sets or different time periods.

[0071] (2) Scale unification: Parameters with different units (such as Celsius and Fahrenheit for temperature, Pascal and Bar for air pressure, etc.) can be converted to the same scale after standardization.

[0072] (3) Data distribution: After standardization, the parameters should meet the requirements of subsequent analysis, for example, they can be normally distributed or uniformly distributed, because some algorithms (such as linear regression, support vector machine, etc.) are highly dependent on the assumptions of data distribution.

[0073] (4) Losslessness: The standardization process should preserve the characteristics of the original data as much as possible to avoid information loss or data deformation.

[0074] (5) Adaptability: The standardized algorithm should be able to adapt to changes in the original data, especially during the production process when parameters may change with changes in feed, equipment status, and other conditions.

[0075] Standardized parameters are an important data processing method that can improve the comparability and consistency of various parameters in the industrial silicon smelting process. Through data cleaning and standardization, various process parameters can be analyzed on a unified scale, providing a basis for subsequent data modeling and optimization. When implementing standardization, conditions such as consistency, scale uniformity, data distribution, non-destructiveness, and adaptability must be met to ensure the quality and validity of the data.

[0076] This step is to systematically obtain and process various key parameters that affect the silica smelting process to ensure the accuracy and consistency of subsequent analysis and optimization processes. In the process of industrial silicon smelting, climate condition parameters such as temperature and humidity, factory environment parameters such as air quality and plant layout, equipment type parameters such as equipment model and operating status, initial raw material parameters such as the initial purity of silica ore and the ingredients and proportions, and reaction process parameters such as temperature, pressure, reaction time, etc. are all important factors that determine the smelting effect. First, the real-time data of these parameters are accurately obtained through the sensor network and data acquisition system. Then, data cleaning is performed to remove noise data and outliers to ensure the accuracy and integrity of the data. In this step, standardization is used to convert parameter data of different dimensions into a unified scale for subsequent analysis and comparison. Standardization not only improves the consistency of the data, but also reduces the impact of dimensional differences between different parameters on model training and prediction. Through this step, a high-quality and reliable data foundation is established, laying a solid foundation for subsequent optimization analysis.

[0077] Step S2: converting the standardized process parameters into weight factors that affect the smelting results, and mapping the relationship between the smelting results and the weight factors into a parameter equation based on a preset multivariate linear regression model, thereby obtaining a parameter set including weight factors and linear regression coefficients.

[0078] This step first identifies the nonlinear relationship that the linear model may not fully capture based on the established linear regression model and regression coefficient. Although linear regression can provide valuable information about the basic effects of each component, there are complex nonlinear effects in the silica smelting process, which requires a more advanced model to characterize. Specifically, the present embodiment selects the exponential regression method to capture the complex nonlinear relationship between the main component and the smelting result, because in the silica smelting process, many reaction rates and efficiencies show an exponential growth or decay trend. For example, the effect of temperature on the chemical reaction rate is usually manifested as an exponential growth, and the reaction rate is significantly accelerated under high temperature conditions, but as the temperature further increases, the acceleration effect gradually weakens. In addition, the influence of humidity and factory environmental parameters on the smelting effect also has obvious nonlinear characteristics, such as humidity. The activity of some reaction reagents has a significant effect, which is manifested as an exponential relationship. After selecting exponential regression, the main component data is first pre-processed to identify the key parameters that have an exponential relationship with the smelting result, and then these parameters are fitted by the exponential regression model to determine the specific relationship between them and the smelting result. The fitting process of the model uses historical data and experimental data, and iterates and adjusts multiple times through the optimization algorithm to finally determine the best fitting parameters. During the regression analysis, corresponding exponential regression models are established for different main components to capture their unique impact on the smelting effect. Finally, through these regression analyses, a set of parameters that can accurately describe the exponential relationship between the main components and the smelting results are obtained, forming a detailed parameter set for subsequent optimization and decision-making. This step ensures that the model can not only reflect the actual situation, but also provide reliable predictions and guidance in a complex and changing production environment, thereby improving the efficiency of the smelting process and product quality.

[0079] As a specific implementation, the step S2 specifically includes the following steps:

[0080] Step S21: extracting main climate condition parameters, main factory environment parameters and main equipment type parameters from the standardized process parameters.

[0081] Step S22: The main climate condition parameters, the main factory environment parameters and the main equipment type parameters are combined into a main component data set; wherein the main component data set includes various main component data.

[0082] Step S23: By comparing the relative importance of the main component data to the smelting results, a weight factor matrix affecting the smelting results is obtained.

[0083] Step S24: performing multivariate linear regression processing on the weight factor matrix, establishing a linear relationship between the main components and the smelting results by inputting the main component data set and the weight factors into a preset regression model, and obtaining a linear regression model and regression coefficients;

[0084] Step S25: performing nonlinear regression analysis processing on the linear regression model and the regression coefficient, and obtaining the parameter set by mapping the nonlinear relationship between the main components and the smelting results into a parameter equation.

[0085] Specifically, the multiple linear regression model in step S24 and step S25 is used to analyze the influence of multiple independent variables (i.e., main components) on a dependent variable (i.e., smelting result). In this process, the model fits the data by the least square method to minimize the sum of squares of errors, thereby finding the best fitting regression coefficient. The goal of the regression model is to make the predicted smelting result as close to the actual smelting result as possible by solving the regression coefficient. By combining the main component data set with the weight factor matrix, the model can accurately reflect the comprehensive influence of each component on the smelting result under different weights, and the calculated regression coefficient represents the linear influence of each main component on the smelting result. This step can quantify the influence of each main component on the smelting result through multiple linear regression processing, and provide a scientific and accurate model to describe the relationship between the main component and the smelting effect. This embodiment adopts the multiple linear regression model in the prior art. The multiple linear regression model (Multiple Linear Regression Model) is a statistical method for analyzing the linear relationship between a dependent variable (response variable) and multiple independent variables (predictor variables). It describes the relationship between these variables by establishing a linear equation, thereby helping to predict and explain the changes in the dependent variable.

[0086] As a more specific implementation, the step S21 specifically includes:

[0087] Performing dimensionality reduction processing on the climate condition parameters to extract the main influencing factors of the climate conditions, thereby obtaining the main climate condition parameters; and,

[0088] Performing dimensionality reduction processing on the factory environment parameters to extract the main influencing factors of the factory environment, thereby obtaining the main factory environment parameters; and,

[0089] The equipment type parameters are subjected to dimensionality reduction processing to extract the main influencing factors of the factory environment, thereby obtaining the main factory environment parameters.

[0090] The core of this step is dimensionality reduction, which converts the original high-dimensional data into low-dimensional data while retaining as much original information as possible. Specifically, multidimensional data such as climate condition parameters, factory environment parameters and equipment type parameters are first calculated through the covariance matrix to calculate their eigenvalues ​​and eigenvectors. The size of the eigenvalue reflects the variance of the data in the direction of the corresponding eigenvector, that is, the amount of information. Then select the eigenvectors with larger eigenvalues ​​as new basis vectors to form a new low-dimensional space. The components corresponding to these basis vectors are called principal components, which are new representations of the original data after dimensionality reduction. Principal component analysis can not only reduce data dimensions and reduce computational complexity, but also effectively eliminate redundant and redundant information, and improve the computational efficiency and stability of the model. Through this method, a main component data set is obtained, which contains the most important change trends and information in the original data. In this embodiment, the main components finally obtained include temperature change trends, humidity change trends, equipment operating status and factory environment parameters. Specifically, temperature is a key parameter affecting the industrial silicon smelting process. Different smelting processes have different temperature requirements. For example, high temperature conditions accelerate the reaction rate, which is conducive to the removal of impurities, but too high a temperature may lead to excessive energy consumption or the generation of by-products; humidity changes can affect the reaction rate and product purity. A high humidity environment may lead to dilution of the reaction reagents, reduce reaction efficiency, and may also affect the operating status of the equipment; different types of equipment and their operating status have a significant impact on the smelting effect. The performance parameters of the equipment, such as furnace temperature stability and stirring rate of the reactor, directly affect the uniformity and efficiency of the smelting process. The layout and environmental conditions of the factory, such as ventilation conditions, dust concentration and pollutant emission levels, will have an impact on silica smelting. For example, good ventilation conditions can effectively reduce the concentration of harmful gases generated during the reaction and reduce the impact on the process and operators. Excessive dust concentration may affect equipment operation and product quality.

[0091] The step S23 specifically includes the following steps:

[0092] The weight factors are calculated according to the main component data sets, and a hierarchical structure model is constructed by applying the hierarchical analysis method. The relative importance of the main component data in each main component data set to the smelting results is compared, and the weight factor of each main component data is calculated to obtain the weight factor matrix affecting the smelting results.

[0093] The calculation formula for the weight factor of each principal component data is as follows:

[0094]

[0095] The expression of the weight factor matrix affecting the smelting result is as follows:

[0096]

[0097] Among them, W i represents the weight factor of the i-th main component data; W is the weight factor matrix; n represents the number of main components; w1 represents the weight factor of the first main component; w2 represents the weight factor of the second main component; w n represents the weight factor of the nth principal component; i and j represent the serial numbers of the principal components; a ij is the element in the importance comparison matrix of the i-th component relative to the j-th component in the hierarchical model.

[0098] This step first builds a hierarchical model based on the extracted main component data set. This model decomposes complex decision-making problems into multiple levels, from the overall goal (smelting results) to each main component (such as temperature, humidity, equipment status, factory environment). When building a hierarchical model, the overall goal is first defined, that is, optimizing smelting efficiency and finished product quality. Then, this goal is decomposed into multiple levels: the first level is the main components, such as temperature, humidity, equipment operating status and factory environmental parameters; the second level is the sub-factors of each specific factor, such as temperature can be further subdivided into the effects of high temperature and low temperature, humidity can be subdivided into the effects of different humidity levels on reaction rate, etc. Preferably, during the silica smelting process, the factory environment and climate conditions have the characteristics of dynamic changes. Environmental dynamic change factors are introduced into the hierarchical model, such as day and night changes in temperature and humidity, seasonal fluctuations, etc., and these changes are analyzed in detail to ensure that the model can adapt to different environmental conditions. After the model is established, the main components in each level are compared pairwise by the paired comparison method to determine their relative importance to the goal. In order to reduce the influence of subjective judgment, an automatic discrimination algorithm is introduced to optimize the process of pairwise comparison through historical data and actual production data. This algorithm is based on data analysis technology and uses a large amount of historical data and actual production data for retrospective analysis, so as to scientifically quantify the importance of each main component. Furthermore, the automatic discrimination algorithm uses machine learning methods to evaluate the smelting effect of each main component under different conditions and generate weight values. For example, through algorithms such as regression analysis or decision trees, the effects of different temperatures, humidity and equipment status on the smelting effect are analyzed to automatically generate a pairwise comparison matrix. In this process, the elements in the matrix represent the importance of each component relative to other components, which are automatically calculated and optimized by the algorithm. The pairwise comparison matrix generated by the automatic discrimination algorithm is checked for consistency to ensure the consistency of judgment. If the consistency index (CI) does not meet the requirements, the comparison matrix is ​​automatically adjusted. Then, the maximum eigenvalue of the matrix and its corresponding eigenvector are calculated, and the eigenvector is normalized to obtain the weight factor of each component. These weight factors constitute the weight factor matrix, which reflects the contribution of each main component in the smelting process.

[0099] As a more specific implementation, the step S25 specifically includes the following steps:

[0100] Performing feature transformation processing on the linear regression model and the regression coefficient, and obtaining a nonlinear data set by performing nonlinear transformation on the main component variables in the linear regression model;

[0101] Performing nonlinear regression processing on the nonlinear data set, recalculating the nonlinear regression coefficient using the nonlinear least square method, thereby establishing a nonlinear regression model between the main components and the smelting results, and obtaining a nonlinear regression equation;

[0102] Performing parameter equation construction processing according to the nonlinear regression equation, expressing the nonlinear relationship in the nonlinear regression equation in the form of a mathematical equation associating the main component variables and the predicted smelting results, and obtaining a preliminary parameter set;

[0103] The preliminary parameter set is subjected to comprehensive evaluation processing, and various parameter equations are integrated to form a comprehensive parameter model to obtain the final parameter set, that is, the corresponding parameter set.

[0104] According to the linear regression model and the regression coefficient, feature transformation processing is performed, and a nonlinear data set is obtained by performing nonlinear transformation on the main component variables in the linear regression model; specifically, based on the established linear regression model, the main component variables are identified, and the potential nonlinear relationship between these variables and the smelting results is considered. The core of feature transformation is to perform appropriate nonlinear processing on each main component variable (such as temperature, humidity, equipment status, etc.) in the linear regression model so that it can more accurately reflect the complex relationship in the actual production process.

[0105] According to the nonlinear data set, nonlinear regression processing is performed, and the nonlinear regression coefficient is recalculated by using the nonlinear least squares method, and a nonlinear regression model between the main component and the smelting result is established to obtain a nonlinear regression equation; specifically, the nonlinear least squares method is used in the nonlinear regression processing because the method can effectively process nonlinear data and find the best fitting curve. The nonlinear least squares method minimizes the square sum of the error between the predicted value and the actual value through iterative optimization, and finally determines the regression coefficient. These regression coefficients will describe the degree of influence of each nonlinear feature on the smelting result. The specific process is as follows: first, the nonlinear data set is input into the nonlinear regression model, and the initial regression coefficient value is set. Then, the nonlinear least squares method is used for multiple iterations, and the error between the predicted result of the model and the actual smelting result is gradually reduced by adjusting the regression coefficient. In each iteration, the optimization algorithm adjusts the regression coefficient according to the current error and gradually approaches the optimal solution. In the iterative process, the model will continuously update the regression coefficient until the preset convergence condition is met, that is, when the error change rate is lower than a certain threshold or reaches the maximum number of iterations, the iteration is stopped. In this way, a regression model that can accurately reflect the nonlinear relationship between the main component and the smelting result is finally established.

[0106] According to the nonlinear regression equation, the parameter equation construction process is performed, and the nonlinear relationship in the nonlinear regression equation is expressed in the form of a mathematical equation that associates the main component variables and the predicted smelting results, and a preliminary parameter set is obtained; specifically, the nonlinear regression equation is first analyzed to identify the specific nonlinear relationship between each main component variable and the smelting result. These relationships are clearly reflected in the form of regression coefficients and nonlinear terms. For example, the exponential growth effect of temperature, the nonlinear effect of humidity, etc., will all be clearly described in mathematical form in the equation. Through this analysis, the specific role of each main component variable in the model can be understood. Next, these analyzed nonlinear relationships are reorganized to construct a set of parameter equations. These parameter equations clarify how the main component variables work together to affect the smelting results. The construction process of the parameter equation needs to consider not only the individual influence of each variable, but also the interaction between multiple variables. Through these parameter equations, the comprehensive influence of the main component variables on the smelting results can be quantitatively described. After completing the construction of the parameter equation, a preliminary parameter set is obtained. This preliminary parameter set contains all the main component variables and their corresponding nonlinear relationships and regression coefficients, providing a comprehensive mathematical model for predicting and optimizing smelting results. The parameter equation is as follows:

[0107]

[0108] Y H =β2H 2 +γ2H;

[0109] Y E =β3E 2 +γ3E;

[0110] Y C =β4ln(C);

[0111] Y=Y T +Y H +Y E +Y C +∈;

[0112] Among them, Y T represents the influence of temperature on the smelting results; β1 is the influence coefficient of temperature, which represents the basic contribution of temperature to the smelting results; α1 is the exponential coefficient of temperature, which represents the exponential growth effect of temperature change on the smelting results; T is the temperature variable; YH is the influence of humidity on the smelting results; β2 is the quadratic term coefficient of humidity, which represents the influence of the square term of humidity on the smelting results; γ2 is the linear term coefficient of humidity, which represents the linear influence of humidity on the smelting results; H is the humidity variable; Y E To reflect the influence of equipment status on smelting results; β3 is the quadratic term coefficient of equipment status, indicating the influence of the square term of equipment status on smelting results; γ3 is the linear term coefficient of equipment status, indicating the linear influence of equipment status on smelting results; E is the equipment status variable; Y C To reflect the influence of factory environmental parameters on smelting results; β4 is the logarithmic coefficient of factory environmental parameters, indicating the logarithmic influence of factory environmental parameters on smelting results; C is the factory environmental parameter variable; Y is the smelting result; ∈ is the error term, indicating the error between the model predicted value and the actual value.

[0113] Among them, the exponential coefficient of temperature can be positive, negative or zero, and the specific value depends on the data analysis and model fitting results in actual operation; the estimation of the linear term coefficient of humidity usually comes from the regression analysis of historical data, that is, by fitting the model, the specific value is obtained according to the smelting results under different humidity conditions; the specific value of the linear term coefficient of equipment status is usually obtained through statistical analysis and model fitting, based on historical data, such as production efficiency, fault records, maintenance cycle, etc.; the value range and specific definition of the equipment status variable depend on the type and scenario of the equipment you are using. For example, common equipment status variables such as equipment efficiency and failure rate have specific value ranges, usually above 0, and the specific range can be adjusted according to actual operation and monitoring standards.

[0114] The plant environmental parameter variable (usually denoted as P) is used in the smelting process to represent various environmental factors that affect the smelting results, such as temperature, humidity, pollutant concentration, etc. The following is a detailed explanation of the plant environmental parameter variable and its value range.

[0115] Factory environmental parameter variables refer to variables related to the environment during the smelting process, including:

[0116] Temperature, that is, the temperature state inside the factory, may affect equipment performance and smelting reactions. The temperature range is generally between -50℃ and 150℃, depending on the actual operating conditions of the factory.

[0117] Humidity, that is, the concentration of water vapor in the air, affects the gas flow and chemical reactions during the smelting process. The humidity value range is usually between 0% and 100%.

[0118] Air quality, including pollutant concentrations (such as dust, sulfur dioxide, etc.), will affect the health of equipment and workers during the smelting process. Its value range is: pollutant concentration is measured in ppm (parts per million), and the specific allowable range depends on regulations and industry standards.

[0119] Vibration, that is, the vibration level in the equipment or factory environment, may affect the operation of the equipment and the stability of the smelting process. Its value range: expressed in acceleration (such as g value) or displacement (such as millimeters), usually ranging from 0 to a certain upper limit (such as 5mm / s, etc.).

[0120] Noise refers to the noise level in the working environment. Its value range is usually expressed in decibels (dB), generally between 30dB and 120dB.

[0121] Air pressure: the air pressure level of the factory environment. Its value range is usually between 500hPa and 1200hPa, depending on the altitude and factory type.

[0122] According to the preliminary parameter set, a comprehensive evaluation process is performed, and the final parameter set is obtained by integrating the various parameter equations to form a comprehensive parameter model, that is, the corresponding parameter set is obtained. Specifically, the comprehensive model combines the nonlinear relationship equations of temperature, humidity, equipment status and factory environmental parameters according to certain weights to form a unified parameter equation. This comprehensive equation not only retains the individual influence of each variable, but also reflects the interaction between them, thereby more accurately describing the overall dynamics of the smelting process.

[0123] Specifically, this step converts parameters from different sources into weight factors that have a direct impact on the silica smelting results, and establishes the relationship between these factors and the smelting effect based on the multiple linear regression model. The weight factors reflect the relative contribution of different parameters to the smelting effect. Next, the multiple linear regression model is used to fit these weight factors to the data of the smelting results to establish a parametric equation that reflects the comprehensive impact of each factor on the smelting result. Through regression analysis, the linear regression coefficients and weight factors obtained constitute a parameter set. This parameter set can not only quantify the impact of each factor, but also provide a specific mathematical expression for predicting and optimizing the smelting effect.

[0124] Step S3: constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters; and, based on the equivalent circuit model, simulating the industrial silicon smelting process using transmission line theory to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables.

[0125] like Figure 2 As shown, as a specific implementation, step S3 specifically includes the following steps:

[0126] Step K1: Mapping the parameter set and the initial raw material parameters to the resistor, capacitor and inductor elements in the circuit model to construct an equivalent circuit model including the initial circuit parameter set.

[0127] It can be understood that the equivalent circuit model is a tool for simulating the actual chemical reaction process, which reflects the influence of various variables in the smelting process through the characteristics of circuit elements. Specifically, the resistance (R), capacitance (C) and inductance (L) elements are used to simulate the characteristics of resistance, reaction rate and reaction delay in the reaction process: resistance (R) represents the reaction resistance, and the linear regression coefficient reflects the resistance effect of certain conditions in the reaction process. These coefficients can be directly mapped to resistance values. Capacitance (C) represents the reaction rate. The purity and component ratio in the initial raw material parameters affect the speed of the reaction. These parameters are mapped into capacitance values, reflecting the influence of different raw material combinations on the reaction speed. Inductance (L) represents reaction delay. The delay effect of certain nonlinear parameters such as temperature on the reaction can be mapped to inductance values, reflecting the dynamic response of the reaction under different temperature conditions. By mapping these linear and nonlinear regression coefficients and the initial raw material parameters to circuit elements, an initial circuit model is constructed. This initial model can simulate the interaction and comprehensive influence of different variables in the silica smelting process, and provides a basis for subsequent parameter optimization. The equivalent circuit model constructed in this embodiment is as follows Figure 2As shown, R1 and R2 represent resistances, and the changes in resistance values ​​represent the energy loss and transmission resistance at different stages in the silica smelting process; C represents capacitance, which is used to reflect the accumulation and release of heat and substances in the silica smelting process; L represents inductance, which is used to simulate the dynamic behavior and delay in the reaction process.

[0128] Step K2: Based on the equivalent circuit model, the heat transfer and mass transfer behaviors in the silica smelting process are simulated by applying the transmission line technology principle, and the voltage and current distribution under different circuit parameter combinations are calculated according to the initial circuit parameter set in combination with the reaction temperature and reaction time to obtain the circuit parameter distribution data.

[0129] It should be noted that the principle of transmission line technology is a theoretical method used to describe the propagation of energy (such as electrical energy, thermal energy, material transfer, etc.) along the transmission path. The heat transfer and mass transfer behaviors in the silica smelting process are mapped to the circuit model. Resistance (R) represents reaction resistance, capacitance (C) represents reaction rate, and inductance (L) represents reaction delay. These circuit elements are connected together through transmission line theory to simulate the transfer and distribution of energy and material in the smelting process. The specific process includes the following steps:

[0130] Step A1, heat transfer simulation: The reaction temperature is used as an input parameter, and the distribution of heat in the system is calculated through transmission line theory to simulate the heat transfer behavior under different temperature conditions. Resistance (R) represents thermal resistance, capacitance (C) represents heat capacity, and inductance (L) represents thermal inertia.

[0131] Step A2, mass transfer simulation: Reaction time is another key parameter, combined with transmission line theory to simulate the transfer of substances in the system. At this time, resistance (R) represents the resistance to material transfer, capacitance (C) represents the speed of material transfer, and inductance (L) represents the delay of material transfer.

[0132] By combining reaction temperature and reaction time, transmission line theory can simulate in detail the distribution of voltage and current under different circuit parameter combinations. The voltage distribution reflects the driving force of energy transfer, while the current distribution represents the rate and delay of material transfer. Through these calculations, the circuit parameter distribution data under different operating conditions and parameter settings can be obtained. The formula is as follows:

[0133]

[0134]

[0135]

[0136] Where t and s represent the time sequence; V(t) represents the voltage distribution at time t; I(s) represents the current at time s; I(t) represents the current distribution at time t; R(s) represents the resistance at time s; C represents the capacitance of the circuit; V0 and I0 represent the initial voltage and initial current; Q(t) represents the accumulated heat at time t; k represents the heat transfer coefficient; T(t) represents the reaction temperature at time t; T amb Indicates the ambient temperature.

[0137] Step K3: Based on the circuit parameter distribution data, the initial circuit parameters are iteratively adjusted using a gradient descent method to obtain an optimized circuit parameter set.

[0138] The circuit parameter distribution data describes in detail the voltage and current distribution under different circuit parameter combinations. Through the circuit parameter distribution data, it is possible to identify which parameter combinations can maximize smelting efficiency, minimize energy consumption, and optimize smelting effects. Next, enter the parameter optimization processing stage. The gradient descent method is an optimization algorithm that iteratively adjusts parameters so that the value of the objective function (such as smelting efficiency, energy consumption, etc.) gradually approaches the optimal value. In this embodiment, the objective function is defined as a loss function that comprehensively considers multiple indicators such as smelting efficiency, impurity removal rate, and energy consumption. The specific optimization process includes the following steps:

[0139] Step B1, initialization parameters: Use the initial values ​​in the circuit parameter distribution data as the starting point of the gradient descent method. These initial values ​​include the initial settings of resistance (R), capacitance (C) and inductance (L).

[0140] Step B2: Calculate the gradient of the objective function for each circuit parameter, that is, the partial derivative of the objective function with respect to the parameter. These gradients represent the rate of change of the objective function under the current parameter settings.

[0141] Step B3: According to the gradient information, iteratively adjust the parameters through the following update rules:

[0142]

[0143] in, θ Indicates circuit parameters (including resistance R, capacitance C, inductance L); θ new represents the updated circuit parameters; θ old Represents the circuit parameters before updating; α represents the learning rate (step size); Represents the gradient of the objective function.

[0144] Step B4: Repeat the process of calculating the gradient and updating the parameters until the value of the objective function converges to the minimum value or reaches the preset number of iterations. The sign of convergence is that the parameter update amplitude gradually decreases and the change amplitude of the objective function is within the preset threshold range.

[0145] Step B5: When the gradient descent method converges, a set of optimized circuit parameters is obtained. These optimized parameters reflect the resistance, capacitance and inductance values ​​that can achieve the best smelting effect under various operating conditions.

[0146] Step K4: By simulating the optimized circuit parameter set, the smelting results under different parameter combinations are obtained and relevant data are generated, thereby obtaining parameter adjustment variables. In this step, the optimized circuit parameter set is first input into the simulation environment, and different parameter combinations are set for simulation. The circuit model used in the simulation environment reflects the key reaction mechanisms in the silica smelting process, including heat transfer, mass transfer behavior, etc. Next, for each parameter combination, the smelting results are calculated. During the simulation process, the voltage distribution and current distribution reflect the energy transfer and material transfer in the smelting process. Through these distributions, key indicators such as smelting efficiency, impurity removal rate, and energy consumption can be evaluated. Then, the calculated smelting results are summarized to generate a set of detailed smelting result data. These data include the smelting effects under various parameter combinations, such as impurity removal rate, silica purity, reaction time, energy consumption, etc. Finally, according to the smelting result data, the parameter combinations that have a significant impact on the smelting effect are identified, and the adjustment variables in these parameter combinations are extracted. The parameter adjustment variables reflect which parameter combinations can optimize the smelting effect under different operating conditions. Through simulation processing and smelting result calculation, the effect of the optimized circuit parameter set in the actual smelting process can be fully evaluated and verified, ensuring that the process can remain efficient and stable under different conditions, thereby improving overall smelting efficiency and product quality.

[0147] Specifically, this step first uses a parameter set to combine the initial raw material parameters and the reaction process parameters to construct an equivalent circuit model. In the equivalent circuit model, various physical and chemical phenomena such as heat transfer, mass transfer and reaction rate correspond to components such as resistance, capacitance and inductance in the circuit. Transmission line theory can simulate the energy transfer and distribution in the smelting process, and reflect the changes in the smelting effect through changes in voltage and current. Specifically, resistance corresponds to reaction resistance, capacitance corresponds to reaction rate, and inductance corresponds to the delay characteristics of the reaction process. By adjusting these circuit parameters, the smelting effect under different combinations of batching parameters can be simulated. Then, different parameter combinations are simulated and calculated to generate a series of smelting result data, which reflect the effects of different parameter combinations on the smelting effect. Finally, based on these result data, parameter adjustment variables are extracted, which will be used for further adjustments in the optimization algorithm.

[0148] Step S4: According to the numerical range of the parameter adjustment variable and the initial raw material parameters, an optimal batching parameter combination is obtained.

[0149] The step of adjusting the numerical range of the parameter variable and the initial raw material parameters to obtain the optimal batching parameter combination specifically includes:

[0150] Step S41: generating a plurality of initial batching parameter combinations according to the numerical range of the parameter adjustment variable and the initial raw material parameters.

[0151] As a specific implementation, step S41 specifically includes the following steps:

[0152] According to the initial raw material parameters, the purity of silica and its ingredients and proportions are analyzed to determine the value range of each parameter, thereby obtaining a parameter range set;

[0153] Based on the parameter range set and parameter adjustment variables, the Latin hypercube sampling method is used to perform non-uniform sampling within the specified range, covering the area with a greater impact on the smelting results with high density, thereby generating a non-uniformly distributed set of sample points, and finally obtaining a preliminary sampling parameter combination set;

[0154] By calculating the reaction driving force and reaction speed of each sampling point in the reaction process in the preliminary sampling parameter combination set, the sampling points that meet the preset optimization target are screened out, thereby obtaining the optimized sampling parameter combination set;

[0155] The optimized sampling parameter combination set is combined and generated to obtain the initial batching parameter combination.

[0156] It should be noted that Latin hypercube sampling is an efficient multidimensional random sampling method that can evenly distribute sampling points in the parameter space, ensuring that all possible values ​​of each parameter can be fully covered, avoiding the uneven sampling problem that may occur in traditional methods. The introduction of the Latin hypercube sampling method makes the sampling process both efficient and representative, and can cover a wider range of parameter combinations with fewer sampling points.

[0157] Specifically, this step includes the following steps:

[0158] Step C1, performing range determination processing according to the initial raw material parameters, determining the value range of each parameter by analyzing the purity of silica and its ingredient composition and proportion in the initial raw material parameters, and obtaining a parameter range set;

[0159] It is understandable that the purity of silica directly affects the difficulty and effect of the smelting process, while different ingredients and proportions affect the reaction rate, energy consumption and by-product generation of the smelting. Through analysis and processing, the value range of each initial raw material parameter is obtained, and these ranges are combined to form a complete parameter range set. This set will be used in subsequent simulation and optimization steps to ensure that all possible parameter combinations are evaluated within a reasonable range.

[0160] Step C2, sampling is performed according to the parameter range set and the parameter adjustment variable, and non-uniform sampling is performed within the parameter range by applying the Latin hypercube sampling method, and the area with greater influence on the smelting result is sampled with a higher density, thereby generating a non-uniformly distributed sample point set and obtaining a preliminary sampling parameter combination set;

[0161] It should be noted that when applying the Latin hypercube sampling method, the emphasis is on non-uniform sampling, that is, higher density sampling of areas that have a greater impact on the smelting results within the parameter range. This method can ensure that all possible value combinations are evenly covered in the parameter space. At the same time, through the analysis of the smelting result data, the parameter areas that have a significant impact on the smelting efficiency, impurity removal rate, energy consumption and product purity are identified. Through this non-uniform sampling strategy, a set of sample points covering a wide range of the parameter space is generated, and each sample point represents a parameter combination. All sample points are aggregated to form a preliminary set of sampling parameter combinations. Through this improved sampling method, a detailed evaluation of key areas is ensured, so that the model can more accurately reflect the complexity of the actual smelting process, improve the robustness and accuracy of the model, and provide reliable sample data for subsequent optimization.

[0162] Step C3, performing optimization processing according to the preliminary sampling parameter combination set, by calculating the reaction driving force and reaction speed of each sampling point in the reaction process, screening out sampling points that meet the preset optimization target, and obtaining an optimized sampling parameter combination set;

[0163] It should be noted that the reaction driving force refers to the energy or force that can drive the reaction during the smelting process. For example, by simulating the chemical reaction kinetics under different temperature and pressure conditions, the reaction driving force corresponding to each sampling point is calculated. The reaction rate refers to the speed of the smelting reaction under given conditions. Specifically, the reaction rate of each sampling point is calculated by simulating the chemical reaction rate equation and combining specific reaction conditions such as temperature and catalyst concentration.

[0164] Step C4: performing combination generation processing according to the optimized sampling parameter combination set to obtain an initial batching parameter combination.

[0165] It is understandable that the parameter sets that meet the conditions are extracted from the optimized sampling parameter combination set. These parameter sets have been screened in the above steps and have the potential to achieve higher smelting efficiency or other optimization goals. By combining these parameters, different initial batching parameter combinations are generated. These parameters include the purity of the silica ore, the types and proportions of various additives, the reaction temperature, the reaction time, etc.

[0166] Step S42: Process the initial batching parameter combinations to obtain the fitness value of each initial batching parameter combination.

[0167] As a specific implementation, step S42 specifically includes the following steps:

[0168] After obtaining the historical data of various parameters and smelting results in the industrial silica smelting process, the fuzzy logic method is used to establish a fuzzy rule base for smelting result prediction, thereby forming a fuzzy reasoning model;

[0169] Input the initial batching parameter combination into the fuzzy reasoning model, and use fuzzy rules and reasoning mechanism to calculate, so as to obtain the predicted value of smelting results under each batching parameter combination;

[0170] Comprehensive evaluation is conducted based on the predicted values ​​of smelting results, which are decomposed into four aspects: impurity removal efficiency, metal silicon purity, gaseous by-product content and energy consumption. The normalization method is used to process the data of each aspect to obtain a comprehensive evaluation index.

[0171] According to the comprehensive evaluation index, the fitness value is calculated using the weighted comprehensive evaluation method. The normalized data of impurity removal efficiency, metallic silicon purity, gaseous by-product content and energy consumption are weighted and summed according to the preset weights to obtain the fitness value of each ingredient parameter combination.

[0172] Specifically, this step first inputs the generated initial batching parameter combination into the fuzzy inference system. Based on fuzzy logic theory, the fuzzy inference system processes the fuzziness and uncertainty of the input parameters through fuzzy rules and membership functions to predict the smelting results. The fuzzy inference method is suitable for processing complex and imprecise data. In the process of silica smelting, it can effectively cope with the influence of environmental changes and process fluctuations. Next, each parameter combination is evaluated according to the predicted smelting results. The evaluation indicators include impurity removal efficiency, that is, the ability to effectively remove impurities during the smelting process; metal silicon purity, that is, the purity of metal silicon in the final product; gaseous by-product content, that is, the content of gaseous by-products such as sulfur dioxide (SO2) and carbon dioxide (CO2) generated during the smelting process; and energy consumption, that is, the energy usage during the entire smelting process. These indicators are normalized to ensure that different indicators are comparable. Finally, the fitness value of each parameter combination is calculated by combining these evaluation indicators. The fitness value reflects the comprehensive performance of the parameter combination in multiple dimensions. A higher fitness value indicates that the parameter combination has a better comprehensive performance in terms of smelting efficiency, product purity, environmental friendliness and energy economy. The fitness value calculation formula is:

[0173] F=w1·X1+w2·X2+w3·(1-X3)+w4·(1-X4);

[0174] Among them, F represents the fitness value; X1 represents the normalized impurity removal efficiency value; X2 represents the normalized metallic silicon purity value; X3 represents the normalized gaseous by-product content value; X4 represents the normalized energy consumption value; w1, w2, w3 and w4 represent the weights of the indicators respectively, satisfying w1+w2+w3+w4=1.

[0175] Specifically, this step includes:

[0176] Step D1, obtaining historical data between various parameters and smelting results in the silica smelting process, applying fuzzy logic methods, establishing a fuzzy rule base for smelting result prediction, and obtaining a fuzzy reasoning model;

[0177] It is understandable that the historical data includes various parameter settings (such as the purity of silica ore, the type and proportion of additives, reaction temperature and time, etc.) and their corresponding smelting results. These data provide the basis for establishing a fuzzy logic model. The fuzzy rule base consists of a series of fuzzy rules that describe the relationship between various input parameters (such as reaction temperature, time, ingredient ratio, etc.) and smelting results. Each rule is expressed in the form of "if...then...", defining how the input conditions affect the output results. For example, "If the reaction temperature is high and the reaction time is long, the smelting efficiency is high." The fuzzy reasoning model finally constructed predicts the smelting results under different ingredient combinations by substituting the parameter values ​​input in the actual operation into the fuzzy rule base for fuzzy reasoning. It can comprehensively consider multiple uncertain factors and give a comprehensive prediction of the smelting results.

[0178] Step D2, inputting the initial batching parameter combination into the fuzzy reasoning model, applying fuzzy rules and fuzzy reasoning engine to perform reasoning calculation, and obtaining the predicted value of smelting result under each batching parameter combination;

[0179] It can be understood that this step is calculated by fuzzy reasoning, and the fuzzy reasoning engine calculates the predicted value of the smelting result for each ingredient combination according to the matching rules and membership. This process includes fuzzification, rule reasoning and defuzzification, and the final output of the predicted value of the smelting result indicates the expected effect under different ingredient settings, such as smelting efficiency and impurity removal rate. This method uses the ability of fuzzy logic to process complex and uncertain data, and can more accurately predict the smelting results under different ingredient combinations, thereby providing a scientific basis for optimizing the combination of ingredient parameters.

[0180] Step D3, performing comprehensive evaluation processing according to the predicted value of the smelting result, by decomposing the predicted value of the smelting result into four aspects: impurity removal efficiency, metal silicon purity, gaseous by-product content and energy consumption, and applying a normalization method to process the data of each aspect to obtain a comprehensive evaluation index;

[0181] It should be noted that the decomposition process involves a detailed analysis of the predicted values ​​to ensure that the smelting effect of each aspect is fully considered. Subsequently, the normalization method is applied to process the data of these aspects to eliminate the impact of different dimensions between the indicators, so that the indicators of each aspect are adjusted to a unified scale. For example, the impurity removal efficiency may be expressed as a percentage, while the energy consumption is expressed in kilojoules. Normalization can convert these data in different units into a standardized range (such as 0 to 1). Through this treatment, a comprehensive evaluation index is obtained, which integrates the performance of various aspects and is used to quantify the comprehensive smelting effect of different combinations of ingredient parameters. This comprehensive evaluation method unifies the effects of different ingredient combinations under one standard.

[0182] Step D4, calculate and process the fitness value according to the comprehensive evaluation index. By applying the weighted comprehensive evaluation method, the normalized data of impurity removal efficiency, metallic silicon purity, gaseous by-product content and energy consumption are weighted and summed according to preset weights to calculate the fitness value of each batching parameter combination.

[0183] It can be understood that the fitness value quantifies the comprehensive effect of different combinations of batching parameters in the smelting process, so that different batching schemes can be effectively compared. In this way, smelting efficiency, product purity, by-product control and energy consumption are comprehensively considered to ensure that the final selected batching parameter combination achieves the best balance in all aspects. The fitness value is the result obtained after weighting the comprehensive evaluation index. Its main purpose is to conduct a comprehensive comprehensive evaluation of the batching parameter combination. The weighted comprehensive evaluation method weights and sums each normalized index according to its relative importance to form a comprehensive evaluation value. The fitness value provides a comprehensive quantitative standard for comparing and selecting the optimal batching parameter combination. The role of this step is to obtain a comprehensive score by integrating the influence of various indicators, so that the final selected batching scheme can achieve the best balance in multiple evaluation dimensions.

[0184] Step S43: Optimizing the initial batching parameter combination according to the fitness value to obtain the optimal batching parameter combination.

[0185] As a specific implementation, step S43 specifically includes the following steps:

[0186] Initialize particles according to the fitness value and the initial batching parameter combination, set each particle to a batching parameter combination, and assign random speed and position, and use the fitness value as the initial fitness value to obtain the initial particle swarm state;

[0187] In the initial particle swarm state, update the speed and position of the particles, adjust the inertia weight, acceleration factor and random factor, and determine the individual optimal position and global optimal position of each particle based on the current fitness value, so as to obtain the updated particle swarm state;

[0188] Evaluate the fitness value according to the updated particle swarm state, and update the individual optimal fitness value and the global optimal fitness value of each particle by comparing the current fitness value of each particle with its initial fitness value, and finally obtain a new fitness value set;

[0189] Iterative optimization is performed according to the new fitness value set. The particle speed and position are updated repeatedly through continuous iteration and the fitness value is evaluated until the preset termination condition is met, thereby searching for the optimal combination of ingredient parameters.

[0190] Specifically, this step can efficiently search for the optimal batching parameter combination in a wide parameter space through the application of the particle swarm optimization algorithm. The dynamic adjustment of the inertia weight and the introduction of the local optimal acceleration mechanism enable the search process to have both global exploration capabilities and rapid convergence to the optimal solution. The optimal batching parameter combination finally obtained has the best performance in terms of smelting efficiency, product purity, environmental friendliness and energy economy, providing an efficient and scientific optimization solution for the silica smelting process.

[0191] Specifically, this step includes:

[0192] Step E1, performing particle initialization processing according to the fitness value and the initial batching parameter combination, by setting each particle to represent a batching parameter combination, and assigning a random speed and position to each particle, and taking the fitness value as the initial fitness value, the initial particle group state is obtained;

[0193] Specifically, first, each particle represents a specific combination of batching parameters, that is, the state of each particle includes specific batching components and proportions. Particles are randomly assigned speed and position during initialization. Such random initialization ensures the diversity of the particle swarm in the search space, thereby preventing the algorithm from falling into a local optimal solution too early. The initial fitness value of each particle is calculated based on the initial batching parameter combination. These values ​​are used to evaluate the quality of the particles and reflect their performance in the smelting process. Through this initialization step, the particle swarm optimization algorithm lays the foundation for the subsequent optimization process, enabling the algorithm to gradually optimize the batching parameter combination based on the initial state and fitness value of the particles to achieve the optimization of the smelting effect.

[0194] Step E2: The initial particle swarm state performs particle position and velocity update processing, updates the velocity and position of each particle by adjusting the inertia weight, acceleration factor and random factor, and determines the individual optimal position and global optimal position of each particle based on the current fitness value to obtain the updated particle swarm state;

[0195] It is understandable that the core of the particle swarm optimization algorithm is to update the state of the particle swarm to improve the optimization effect. First, the speed and position of the particles are updated by adjusting the inertia weight, acceleration factor and random factor. The inertia weight determines the impact of the particle's previous speed on the current speed. A higher inertia weight helps to explore a wider search space, while a lower weight helps the particles focus on the current optimal area. Considering the complexity and variability of the silica smelting process, the inertia weight is set high in the initial stage to encourage particles to search the possible solution space extensively; as the optimization process deepens, the inertia weight is gradually reduced to enhance the particle's search ability near the local optimal solution. The acceleration factor controls the speed at which the particles move to the individual optimal position and the global optimal position, which can ensure that the particle swarm converges to the optimal solution. The random factor introduces randomness to avoid the algorithm from falling into the local optimal solution and ensure a more comprehensive search. The speed and position of the particles will be recalculated according to the current fitness value during the update process to determine the individual optimal position of each particle and the global optimal position of the entire particle swarm. Through this update process, the state of the particle swarm will be continuously adjusted, gradually approaching the optimal combination of batching parameters, and improving the optimization effect of the silica smelting process. The specific update rules are as follows:

[0196] The speed update of particles consists of three parts: inertia weight, individual optimal position and global optimal position. The inertia weight controls the continuity of the particle's current speed, the individual optimal position guides the particle to move to the best position it has experienced, and the global optimal position guides the particle to move to the best position experienced by the entire population. The speed update formula is:

[0197]

[0198] Among them, z represents the particle number; t represents the time number; v z (t) represents the velocity of particle z at time t; v z (t+1) represents the speed of particle z at (t+1); ω represents the inertia weight; x z (t) represents the position of particle z at time t; c1 and c2 represent acceleration constants; r1 and r2 represent random numbers, generated in the range of [0, 1], which are used to introduce randomness and prevent particles from falling into local optimality; represents the individual optimal position of particle z; G best Represents all optimal positions.

[0199] The particle position is updated based on the current position and the updated velocity. The position update formula is:

[0200] X z (t+1)=x z (t)+v z (t+1);

[0201] Among them, x z (t+1) represents the position of particle z at (t+1); x z (t) represents the position of particle z at time t; v z (t+1) represents the velocity of particle z at (t+1).

[0202] Step E3, performing fitness value evaluation processing according to the updated particle swarm state, by comparing the current fitness value of each particle with its initial fitness value, updating the individual optimal fitness value and the global optimal fitness value of each particle, and obtaining a new fitness value set;

[0203] Specifically, the current fitness value of each particle is compared with its initial fitness value to evaluate the performance improvement of the particle. Through this comparison, it is possible to identify which particles perform better in the current iteration, and then update their individual optimal fitness values. If the current fitness value of a particle is better than its historical best fitness value, the new fitness value of the particle will be recorded as its new individual optimal fitness value. At the same time, the global optimal fitness value of the entire particle swarm may also be updated, that is, if the fitness value of a particle is better than the currently recorded global optimal fitness value, then the global optimal fitness value and the corresponding particle position will be updated. This evaluation and update process ensures that the particle swarm can continue to approach the optimal solution, thereby effectively optimizing the batching parameters in the silica smelting process.

[0204] Step E4: perform iterative optimization processing according to the new fitness value set, and through continuous iteration, repeatedly update the speed and position of the particles and evaluate the fitness value until the preset termination condition is met, and the optimal batching parameter combination is searched.

[0205] It should be noted that the particle speed and position are continuously improved through repeated iterations, so as to gradually approach the optimal solution. In each iteration, the particle speed and position are updated according to the latest fitness value, and the fitness value of each particle is re-evaluated. This process will continue until the preset termination condition is met, such as reaching the maximum number of iterations, the fitness value change is less than a certain threshold, or the optimization result meets a certain standard. Through this continuous optimization process, the particle swarm can continuously explore and optimize the batching parameters, and finally find the optimal batching parameter combination to improve the effect of the silica smelting process.

[0206] Step S5: optimizing the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

[0207] This embodiment comprehensively considers multiple factors such as factory environment, climate conditions and equipment type, and uses mathematical models and intelligent algorithms to optimize the ingredient parameters. It can stably achieve precise control and rapid optimization and adjustment of the ingredient parameters during the industrial silicon smelting process, thereby more effectively responding to the complex and changeable environment and process requirements in industrial silicon smelting.

[0208] Embodiment 2:

[0209] like Figure 3 As shown, this embodiment provides a system for optimizing batching parameters in an industrial silicon smelting process, the system comprising:

[0210] The acquisition module 10 is used to acquire the process parameters in the industrial silicon smelting process; the process parameters include climate condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters;

[0211] A standardization module 20, connected to the acquisition module 10, for performing standardization processing on the process parameters;

[0212] The first processing module 30 is connected to the standardization module 20 and is used to convert the standardized process parameters into weight factors that affect the smelting results, and based on a preset multivariate linear regression model, map the relationship between the smelting results and the weight factors into a parameter equation, thereby obtaining a parameter set including weight factors and linear regression coefficients;

[0213] A construction module 40, connected to the first processing module 30, for constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters;

[0214] A generating module 50, connected to the building module 40, is used to simulate the industrial silicon smelting process based on the equivalent circuit model to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables;

[0215] The optimization module 60 is connected to the generation module 50 and is used to adjust the value range of the variable and the initial raw material parameters according to the parameters to obtain the optimal batching parameter combination;

[0216] The second processing module 70 is connected to the optimization module 60 and is used to optimize the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

[0217] As a specific implementation, the first processing module 30 includes:

[0218] An extraction unit, used to extract main climate condition parameters, main factory environment parameters and main equipment type parameters from the standardized process parameters;

[0219] a combining unit, connected to the extracting unit, for combining the main climate condition parameters, the main factory environment parameters and the main equipment type parameters into a main component data set;

[0220] Wherein, the main component data set includes each main component data;

[0221] A comparison unit, connected to the combination unit, for obtaining a weight factor matrix affecting the smelting result by comparing the relative importance of the main component data to the smelting result;

[0222] A first construction unit, connected to the comparison unit, is used to perform a multivariate linear regression process on the weight factor matrix, establish a linear relationship between the main components and the smelting results by inputting the main component data set and the weight factor into a preset regression model, and obtain a linear regression model and regression coefficients;

[0223] A mapping unit is connected to the first construction unit and is used to perform nonlinear regression analysis based on the linear regression model and regression coefficients, and obtain the parameter set by mapping the nonlinear relationship between the main components and the smelting results into a parameter equation.

[0224] As a specific implementation, the building block 40 includes:

[0225] A second construction unit is used to map the parameter set and the initial raw material parameters to the resistance, capacitance and inductance elements in the circuit model to construct an equivalent circuit model including the initial circuit parameter set;

[0226] The generation module comprises:

[0227] A first calculation unit is connected to the second construction unit, and is used to simulate the heat transfer and mass transfer behaviors in the silica smelting process based on the equivalent circuit model by applying the transmission line technology principle, and calculate the voltage and current distribution under different circuit parameter combinations according to the initial circuit parameter set in combination with the reaction temperature and reaction time, so as to obtain circuit parameter distribution data;

[0228] A first iteration unit, connected to the first calculation unit, is used to iteratively adjust the initial circuit parameters based on the circuit parameter distribution data by using a gradient descent method to obtain an optimized circuit parameter set;

[0229] The first generating unit is connected to the first iterative unit and is used to obtain smelting results under different parameter combinations and generate relevant data by simulating the optimized circuit parameter set, thereby obtaining parameter adjustment variables.

[0230] As a specific implementation, the optimization module includes:

[0231] A second generating unit, configured to generate a plurality of initial batching parameter combinations according to the numerical range of the parameter adjustment variable and the initial raw material parameters;

[0232] a processing unit connected to the second generating unit, and configured to process the initial batching parameter combinations to obtain a fitness value of each initial batching parameter combination;

[0233] The optimization unit is connected to the processing unit and is used to optimize the initial batching parameter combination according to the fitness value to obtain the optimal batching parameter combination.

[0234] As a specific implementation manner, the second generating unit includes:

[0235] The analysis subunit is used to analyze the purity of silica and its ingredients and proportions according to the initial raw material parameters to determine the value range of each parameter, thereby obtaining a parameter range set;

[0236] A generating subunit connected to the analyzing subunit is used to perform non-uniform sampling based on the parameter range set and the numerical range of the parameter adjustment variable by using the Latin hypercube sampling method, so as to cover the area with a large impact on the smelting result with high density, thereby generating a non-uniformly distributed sample point set, and finally obtaining a preliminary sampling parameter combination set; A screening subunit is connected to the generating subunit, and is used to screen out sampling points that meet the preset optimization target by calculating the reaction driving force and reaction speed of each sampling point in the preliminary sampling parameter combination set during the reaction process, thereby obtaining an optimized sampling parameter combination set;

[0237] The combination subunit is connected to the screening subunit and is used to combine and generate the optimized sampling parameter combination set to obtain an initial batching parameter combination.

[0238] As a specific implementation, the processing unit includes:

[0239] Constructing subunits to obtain historical data of various parameters and smelting results in the industrial silica smelting process, and then applying fuzzy logic methods to establish a fuzzy rule base for smelting result prediction, thereby forming a fuzzy reasoning model;

[0240] A calculation subunit, connected to the construction subunit, is used to input the initial batching parameter combination into the fuzzy reasoning model, and use fuzzy rules and reasoning mechanisms to perform calculations, thereby obtaining a predicted value of the smelting result under each batching parameter combination;

[0241] The decomposition subunit is connected to the calculation subunit and is used to make a comprehensive evaluation based on the predicted value of the smelting result, decomposing it into four aspects: impurity removal efficiency, metal silicon purity, gaseous by-product content and energy consumption, and applying the normalization method to process the data of each aspect to obtain a comprehensive evaluation index;

[0242] The weighted subunit is connected to the decomposition subunit and is used to calculate the fitness value according to the comprehensive evaluation index by applying the weighted comprehensive evaluation method, and the normalized data of impurity removal efficiency, metallic silicon purity, gaseous by-product content and energy consumption are weighted and summed according to preset weights to obtain the fitness value of each ingredient parameter combination.

[0243] As a specific implementation, the optimization unit includes:

[0244] The initialization subunit is used to initialize particles according to the fitness value and the initial batching parameter combination, setting each particle to a batching parameter combination, and assigning a random speed and position, while taking the fitness value as the initial fitness value, thereby obtaining the initial particle group state;

[0245] An updating subunit, connected to the initialization subunit, is used to update the speed and position of the particles in the initial particle swarm state, and to obtain an updated particle swarm state by adjusting the inertia weight, acceleration factor and random factor, and determining the individual optimal position and the global optimal position of each particle based on the current fitness value;

[0246] An optimization subunit, connected to the update subunit, for evaluating the fitness value according to the updated particle swarm state, and updating the individual optimal fitness value and the global optimal fitness value of each particle by comparing the current fitness value of each particle with its initial fitness value, and finally obtaining a new fitness value set;

[0247] The iterative subunit is connected to the optimization subunit and is used to perform iterative optimization according to the new fitness value set, repeatedly updating the speed and position of the particles and evaluating the fitness value through continuous iteration until the preset termination condition is met, thereby searching for the optimal batching parameter combination.

[0248] The system in this embodiment can execute the method in Embodiment 1.

[0249] Embodiment 3:

[0250] This embodiment provides an industrial silicon production equipment, characterized in that the industrial silicon production equipment includes the system for optimizing batching parameters in the industrial silicon smelting process described in Example 2.

[0251] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for optimizing batching parameters in an industrial silicon smelting process, characterized in that: The method comprises the following steps: Step S1: obtaining process parameters in an industrial silicon smelting process; and standardizing the process parameters; The process parameters include climatic condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters; Step S2: converting the standardized process parameters into weight factors that affect the smelting results, and mapping the relationship between the smelting results and the weight factors into a parameter equation based on a preset multivariate linear regression model, thereby obtaining a parameter set including weight factors and linear regression coefficients; Step S3: constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters; and simulating the industrial silicon smelting process based on the equivalent circuit model to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables; Step S4: according to the numerical range of the parameter adjustment variable and the initial raw material parameters, an optimal batching parameter combination is obtained; Step S5: optimizing the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

2. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 1, characterized in that: The step S2 specifically includes the following steps: Step S21: extracting main climate condition parameters, main factory environment parameters and main equipment type parameters from the standardized process parameters; Step S22: The main climate condition parameters, the main factory environment parameters and the main equipment type parameters are combined into a main component data set; Wherein, the main component data set includes each main component data; Step S23: by comparing the relative importance of the main component data to the smelting results, a weight factor matrix affecting the smelting results is obtained; Step S24: performing multivariate linear regression processing on the weight factor matrix, establishing a linear relationship between the main components and the smelting results by inputting the main component data set and the weight factors into a preset regression model, and obtaining a linear regression model and regression coefficients; Step S25: performing nonlinear regression analysis on the linear regression model and the regression coefficients, and obtaining the parameter set by mapping the nonlinear relationship between the main components and the smelting results into a parameter equation.

3. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 2, characterized in that: The step S21 specifically includes: Performing dimensionality reduction processing on the climate condition parameters to extract the main influencing factors of the climate conditions, thereby obtaining the main climate condition parameters; and, Performing dimensionality reduction processing on the factory environment parameters to extract the main influencing factors of the factory environment, thereby obtaining the main factory environment parameters; and, Performing dimensionality reduction processing on the equipment type parameters to extract the main influencing factors of the factory environment, thereby obtaining the main factory environment parameters; The step S23 specifically includes the following steps: The weight factors are calculated according to the main component data sets, and a hierarchical structure model is constructed by applying the hierarchical analysis method. The relative importance of the main component data in each main component data set to the smelting results is compared, and the weight factor of each main component data is calculated to obtain the weight factor matrix affecting the smelting results. The calculation formula for the weight factor of each principal component data is as follows: The expression of the weight factor matrix affecting the smelting result is as follows: Among them, W i represents the weight factor of the i-th principal component data; W represents the weight factor matrix; n represents the number of principal components; w1 represents the weight factor of the first principal component; w2 represents the weight factor of the second principal component; ..., w n represents the weight factor of the nth principal component; i and j represent the serial numbers of the principal components; a ij is the element in the importance comparison matrix of the i-th component relative to the j-th component in the hierarchical model.

4. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 2, characterized in that: The step S25 specifically includes the following steps: Performing feature transformation processing on the linear regression model and the regression coefficient, and obtaining a nonlinear data set by performing nonlinear transformation on the main component variables in the linear regression model; Performing nonlinear regression processing on the nonlinear data set, recalculating the nonlinear regression coefficient using the nonlinear least square method, thereby establishing a nonlinear regression model between the main components and the smelting results, and obtaining a nonlinear regression equation; Performing parameter equation construction processing according to the nonlinear regression equation, expressing the nonlinear relationship in the nonlinear regression equation in the form of a mathematical equation associating the main component variables and the predicted smelting results, and obtaining a preliminary parameter set; The preliminary parameter set is subjected to comprehensive evaluation processing, and various parameter equations are integrated to form a comprehensive parameter model to obtain the final parameter set, that is, the corresponding parameter set.

5. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 1, characterized in that: The step S3 specifically includes the following steps: Mapping the parameter set and the initial raw material parameters to resistance, capacitance and inductance elements in the circuit model to construct an equivalent circuit model including the initial circuit parameter set; Based on the equivalent circuit model, the heat transfer and mass transfer behaviors in the silica smelting process are simulated by applying the transmission line technology principle, and the voltage and current distribution under different circuit parameter combinations are calculated according to the initial circuit parameter set in combination with the reaction temperature and reaction time to obtain circuit parameter distribution data; Based on the circuit parameter distribution data, the initial circuit parameters are iteratively adjusted using a gradient descent method to obtain an optimized circuit parameter set; By simulating the optimized circuit parameter set, the smelting results under different parameter combinations are obtained and relevant data are generated, thereby obtaining parameter adjustment variables.

6. The method for optimizing batching parameters in an industrial silicon smelting process according to any one of claims 1 to 5, characterized in that: The step of adjusting the numerical range of the parameter variable and the initial raw material parameters to obtain the optimal batching parameter combination specifically includes: Step S41: generating a plurality of initial batching parameter combinations according to the numerical range of the parameter adjustment variable and the initial raw material parameters; Step S42: processing the initial batching parameter combination to obtain a fitness value of each initial batching parameter combination; Step S43: Optimizing the initial batching parameter combination according to the fitness value to obtain the optimal batching parameter combination.

7. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 6, characterized in that: The step S41 specifically includes the following steps: According to the initial raw material parameters, the purity of silica and its ingredients and proportions are analyzed to determine the value range of each initial raw material parameter, thereby obtaining a parameter range set; Based on the parameter range set and the numerical range of the parameter adjustment variable, a Latin hypercube sampling method is used to perform non-uniform sampling, so as to cover the area with a greater impact on the smelting result with high density, thereby generating a non-uniformly distributed sample point set, and finally obtaining a preliminary sampling parameter combination set; By calculating the reaction driving force and reaction speed of each sampling point in the reaction process in the preliminary sampling parameter combination set, the sampling points that meet the preset optimization target are screened out, thereby obtaining the optimized sampling parameter combination set; The optimized sampling parameter combination set is combined and generated to obtain the initial batching parameter combination.

8. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 6, characterized in that: The step S42 specifically includes the following steps: The initial batching parameter combination is input into the fuzzy inference model, and the fuzzy rules and inference mechanism are used for calculation, so as to obtain the predicted value of the smelting result under each batching parameter combination; wherein the fuzzy inference model is formed by obtaining the historical data of various parameters and smelting results in the industrial silica smelting process, and applying the fuzzy logic method to establish a fuzzy rule library for smelting result prediction; Comprehensively evaluate the predicted value of smelting results, decompose it into four aspects: impurity removal efficiency, metal silicon purity, gaseous by-product content and energy consumption, and apply normalization method to process data from various aspects to obtain comprehensive evaluation indicators; According to the comprehensive evaluation index, the fitness value is calculated using the weighted comprehensive evaluation method. The normalized data of impurity removal efficiency, metallic silicon purity, gaseous by-product content and energy consumption are weighted and summed according to the preset weights to obtain the fitness value of each ingredient parameter combination.

9. The method for optimizing batching parameters in the process of industrial silicon smelting according to claim 6, characterized in that: The step S43 specifically includes the following steps: Initialize particles according to the fitness value and the initial batching parameter combination, set each particle to a batching parameter combination, and assign random speed and position, and use the fitness value as the initial fitness value to obtain the initial particle swarm state; In the initial particle swarm state, update the speed and position of the particles, adjust the inertia weight, acceleration factor and random factor, and determine the individual optimal position and global optimal position of each particle based on the current fitness value, so as to obtain the updated particle swarm state; Evaluate the fitness value according to the updated particle swarm state, and update the individual optimal fitness value and the global optimal fitness value of each particle by comparing the current fitness value of each particle with its initial fitness value, and finally obtain a new fitness value set; Iterative optimization is performed according to the new fitness value set. The particle speed and position are updated repeatedly through continuous iteration and the fitness value is evaluated until the preset termination condition is met, thereby searching for the optimal combination of ingredient parameters.

10. A system for optimizing batching parameters in an industrial silicon smelting process, characterized in that: The system comprises: An acquisition module is used to acquire process parameters in the industrial silicon smelting process; the process parameters include climate condition parameters, factory environment parameters, equipment type parameters, initial raw material parameters and reaction process parameters; A standardization module, connected to the acquisition module, for performing standardization processing on the process parameters; A first processing module, connected to the standardization module, is used to convert the standardized process parameters into weight factors that affect the smelting results, and based on a preset multivariate linear regression model, map the relationship between the smelting results and the weight factors into a parameter equation, thereby obtaining a parameter set including weight factors and linear regression coefficients; A construction module, connected to the first processing module, for constructing an equivalent circuit model based on the parameter set, the initial raw material parameters and the reaction process parameters; A generating module, connected to the building module, for simulating the industrial silicon smelting process based on the equivalent circuit model to generate smelting result data under different parameter combinations, thereby obtaining parameter adjustment variables; An optimization module, connected to the generation module, for adjusting the numerical range of the variables and the initial raw material parameters according to the parameters to obtain an optimal batching parameter combination; The second processing module is connected to the optimization module and is used to optimize the batching parameters in the industrial silicon smelting process according to the optimal batching parameter combination.

11. The system for optimizing batching parameters in the process of industrial silicon smelting according to claim 10, characterized in that: The first processing module comprises: An extraction unit, used to extract main climate condition parameters, main factory environment parameters and main equipment type parameters from the standardized process parameters; a combining unit, connected to the extracting unit, for combining the main climate condition parameters, the main factory environment parameters and the main equipment type parameters into a main component data set; Wherein, the main component data set includes each main component data; A comparison unit, connected to the combination unit, for obtaining a weight factor matrix affecting the smelting result by comparing the relative importance of the main component data to the smelting result; A first construction unit, connected to the comparison unit, is used to perform a multivariate linear regression process on the weight factor matrix, establish a linear relationship between the main components and the smelting results by inputting the main component data set and the weight factor into a preset regression model, and obtain a linear regression model and regression coefficients; A mapping unit is connected to the first construction unit and is used to perform nonlinear regression analysis based on the linear regression model and regression coefficients, and obtain the parameter set by mapping the nonlinear relationship between the main components and the smelting results into a parameter equation.

12. The system for optimizing batching parameters in the process of industrial silicon smelting according to claim 10, characterized in that: The building blocks include: A second construction unit is used to map the parameter set and the initial raw material parameters to the resistance, capacitance and inductance elements in the circuit model to construct an equivalent circuit model including the initial circuit parameter set; The generation module comprises: A first calculation unit is connected to the second construction unit, and is used to simulate the heat transfer and mass transfer behaviors in the silica smelting process based on the equivalent circuit model by applying the transmission line technology principle, and calculate the voltage and current distribution under different circuit parameter combinations according to the initial circuit parameter set in combination with the reaction temperature and reaction time, so as to obtain circuit parameter distribution data; A first iteration unit, connected to the first calculation unit, is used to iteratively adjust the initial circuit parameters by using a gradient descent method based on the circuit parameter distribution data to obtain an optimized circuit parameter set; The first generating unit is connected to the first iterative unit and is used to obtain smelting results under different parameter combinations and generate relevant data by simulating the optimized circuit parameter set, thereby obtaining parameter adjustment variables.

13. The system for optimizing batching parameters in an industrial silicon smelting process according to any one of claims 10 to 12, characterized in that: The optimization module includes: A second generating unit, configured to generate a plurality of initial batching parameter combinations according to the numerical range of the parameter adjustment variable and the initial raw material parameters; a processing unit connected to the second generating unit, and configured to process the initial batching parameter combinations to obtain a fitness value of each initial batching parameter combination; The optimization unit is connected to the processing unit and is used to optimize the initial batching parameter combination according to the fitness value to obtain the optimal batching parameter combination.

14. The system for optimizing batching parameters in the process of industrial silicon smelting according to claim 13, characterized in that: The second generating unit comprises: The analysis subunit is used to analyze the purity of silica and its ingredients and proportions according to the initial raw material parameters to determine the value range of each parameter, thereby obtaining a parameter range set; A generating subunit connected to the analyzing subunit is used to perform non-uniform sampling based on the parameter range set and the numerical range of the parameter adjustment variable by using the Latin hypercube sampling method, so as to cover the area with a large impact on the smelting result with high density, thereby generating a non-uniformly distributed sample point set, and finally obtaining a preliminary sampling parameter combination set; A screening subunit is connected to the generating subunit, and is used to screen out sampling points that meet the preset optimization target by calculating the reaction driving force and reaction speed of each sampling point in the preliminary sampling parameter combination set during the reaction process, thereby obtaining an optimized sampling parameter combination set; The combination subunit is connected to the screening subunit and is used to combine and generate the optimized sampling parameter combination set to obtain an initial batching parameter combination.

15. The system for optimizing batching parameters in the process of industrial silicon smelting according to claim 13, characterized in that: The processing unit comprises: Constructing subunits to obtain historical data of various parameters and smelting results in the industrial silica smelting process, and then applying fuzzy logic methods to establish a fuzzy rule base for smelting result prediction, thereby forming a fuzzy reasoning model; A calculation subunit, connected to the construction subunit, is used to input the initial batching parameter combination into the fuzzy reasoning model, and use fuzzy rules and reasoning mechanisms to perform calculations, thereby obtaining a predicted value of the smelting result under each batching parameter combination; The decomposition subunit is connected to the calculation subunit and is used to make a comprehensive evaluation based on the predicted value of the smelting result, decomposing it into four aspects: impurity removal efficiency, metal silicon purity, gaseous by-product content and energy consumption, and applying the normalization method to process the data of each aspect to obtain a comprehensive evaluation index; The weighted subunit is connected to the decomposition subunit and is used to calculate the fitness value according to the comprehensive evaluation index by applying the weighted comprehensive evaluation method, and the normalized data of impurity removal efficiency, metallic silicon purity, gaseous by-product content and energy consumption are weighted and summed according to preset weights to obtain the fitness value of each ingredient parameter combination.

16. The system for optimizing batching parameters in the process of industrial silicon smelting according to claim 13, characterized in that: The optimization unit comprises: The initialization subunit is used to initialize particles according to the fitness value and the initial batching parameter combination, setting each particle to a batching parameter combination, and assigning a random speed and position, while taking the fitness value as the initial fitness value, thereby obtaining the initial particle group state; An updating subunit, connected to the initialization subunit, is used to update the speed and position of the particles in the initial particle swarm state, and to obtain an updated particle swarm state by adjusting the inertia weight, acceleration factor and random factor, and determining the individual optimal position and the global optimal position of each particle based on the current fitness value; An optimization subunit, connected to the update subunit, for evaluating the fitness value according to the updated particle swarm state, and updating the individual optimal fitness value and the global optimal fitness value of each particle by comparing the current fitness value of each particle with its initial fitness value, and finally obtaining a new fitness value set; The iterative subunit is connected to the optimization subunit and is used to perform iterative optimization according to the new fitness value set, repeatedly updating the speed and position of the particles and evaluating the fitness value through continuous iteration until the preset termination condition is met, thereby searching for the optimal batching parameter combination.

17. An industrial silicon production equipment, characterized in that: The industrial silicon production equipment comprises the system for optimizing batching parameters in the industrial silicon smelting process as described in any one of claims 10 to 16.

Images