Industrial silicon smelting furnace control method and system based on infrared image analysis

Abstract

The present invention discloses an industrial silicon smelting furnace control method and system based on infrared image analysis, which relates to the technical field of industrial silicon smelting furnace control. The method and system analyze the smelting characteristics in the smelting furnace, construct a virtual three-dimensional model, and add an infrared representation layer to the model in combination with the infrared thermal imaging image, and then perform grayscale processing; by analyzing the characteristics of the grayscale layer, a grayscale representation extension model in the furnace is constructed, and its dynamic changes are captured; based on the equality of grayscale value, dynamic representation and historical control parameter data, the model is screened and classified multiple times to form a number of virtual smelting furnace three-dimensional model dynamic representation sets; in the later smelting furnace control, key factors are used to match the current furnace condition with the dynamic representation set, and virtual models with equivalent representation are screened out, and the historical control parameters are used as a reference to guide the real-time control of the smelting furnace. The present invention realizes accurate and efficient smelting process management through the above technical scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial silicon smelting furnace control, and in particular to an industrial silicon smelting furnace control method and system based on infrared image analysis. Background Art

[0002] In the process of industrial silicon smelting, accurate monitoring and control of the surface temperature of the workpiece in the furnace is crucial. Traditional temperature measurement methods are mostly invasive or contact methods, which have many shortcomings, such as limited measurement range, complex operation and easy damage to the workpiece surface. With the continuous development of infrared thermal imaging technology, non-contact temperature monitoring methods based on infrared image analysis have gradually become a research hotspot.

[0003] Infrared thermal imaging technology can intuitively display the temperature distribution of an object by capturing infrared radiation on the surface of the object and converting it into a visual image. This technology not only has the advantages of real-time monitoring, non-contact measurement, full coverage and high precision, but can also work stably in harsh environments such as high temperature, high pressure and high speed operation. In industrial production, infrared thermal imaging technology has been widely used in furnace tube temperature measurement, equipment online fault diagnosis and safety protection of industrial silicon smelting furnaces, significantly improving production efficiency and safety.

[0004] However, the existing infrared thermal imaging technology still faces some challenges in practical applications. For example, how to accurately extract temperature information from infrared images and conduct effective processing and analysis; how to design reasonable control methods and systems based on the actual working conditions of industrial silicon smelting furnaces; how to ensure the stability and reliability of the system to meet the high-precision and high-efficiency requirements in the industrial silicon smelting process, etc. Summary of the invention

[0005] The object of the present invention is to provide a control method and system capable of improving the control accuracy and stability of an industrial silicon smelting furnace.

[0006] The present invention discloses an industrial silicon smelting furnace control method based on infrared image analysis, comprising:

[0007] Collect historical smelting furnace control parameter data, collect historical infrared thermal imaging image records at the same time, and collect simultaneous smelting records. Based on the analysis of the smelting records, eliminate the historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results;

[0008] The design structure image of the smelting furnace is analyzed, and based on the analysis result, a virtual smelting furnace three-dimensional model is constructed, and based on the analysis of several historical infrared thermal imaging images at the same time, an infrared expression layer is configured on the side of the virtual smelting furnace three-dimensional model, and the infrared expression layer is grayed to obtain a gray expression layer;

[0009] The grayscale value characteristics of the grayscale expression layer are analyzed to determine the grayscale value characteristics of different blocks, and the grayscale value characteristics of the virtual smelting furnace three-dimensional model are analyzed for correlation. Based on the analysis results, the furnace grayscale expression extension model is constructed, and the dynamic expression of the furnace grayscale expression extension model is constructed according to the temporal change of the furnace grayscale expression extension model.

[0010] Based on the equivalence between the grayscale value features of the grayscale representation layers, the dynamic representation of the grayscale representation extension model in the furnace is first screened and classified, and based on the equivalence between the dynamic representations of the grayscale representation extension model in the furnace, the dynamic representation of the grayscale representation extension model in the furnace is secondly screened and classified, and based on the equivalence between the historical smelting furnace control parameter data, the dynamic representation of the grayscale representation extension model in the furnace that is screened and classified for the second time is thirdly screened and classified to obtain a number of virtual smelting furnace three-dimensional model dynamic representation sets;

[0011] A key factor analysis is performed on the dynamic performance set of each virtual smelting furnace three-dimensional model. When the smelting furnace is controlled in the later stage, the matching between the key factors and the dynamic performance of the grayscale performance extension model in the current furnace constructed for the current smelting furnace is used to determine the dynamic performance set of the virtual smelting furnace three-dimensional model to be called, and the virtual smelting furnace three-dimensional model with the same dynamic performance as the grayscale performance extension model in the current furnace is screened out from the dynamic performance set of the virtual smelting furnace three-dimensional model, and the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model provides a reference for the current control parameters of the smelting furnace.

[0012] In some embodiments disclosed in the present invention, a method for configuring an infrared display layer on the side of a virtual smelting furnace three-dimensional model includes:

[0013] A positioning array is set for each historical infrared thermal imaging image, the positioning array includes a number of evenly arranged positioning points, and a corresponding positioning block is set for each positioning point, and each positioning point corresponds to a preset position on the side of the virtual smelting furnace three-dimensional model;

[0014] The infrared thermal imaging block corresponding to the positioning block on the historical infrared thermal imaging image is intercepted, and the intercepted infrared thermal imaging block is mapped to the side of the virtual smelting furnace three-dimensional model.

[0015] In some embodiments disclosed in the present invention, the method for performing grayscale value feature analysis on a grayscale representation layer includes:

[0016] Establishing a number of circle analysis lines for the grayscale expression layer, and setting a number of grayscale detection points symmetrically above and below the circle analysis line, collecting the grayscale value of each grayscale detection point, and calculating the grayscale value difference vector between adjacent grayscale detection points;

[0017] Perform continuity analysis on the grayscale value difference vectors. If the change directions of adjacent grayscale value difference vectors are the same and the values ​​of the grayscale value difference vectors are less than or equal to a preset value, then associate the corresponding grayscale detection points until a grayscale detection point cluster is formed.

[0018] Calculate the distribution ratio of each grayscale detection point cluster above and below different circle analysis lines to obtain a distribution ratio group, and associate the distribution ratio group with the grayscale detection point cluster;

[0019] The number of grayscale detection points contained in each grayscale detection point cluster is analyzed, and the average grayscale value of the grayscale detection point cluster is calculated, and the distribution ratio group, the number of grayscale detection points and the average grayscale value are identified as grayscale value features.

[0020] In some embodiments disclosed in the present invention, a method for performing correlation analysis on grayscale value features of a virtual smelting furnace three-dimensional model includes:

[0021] Based on the difference in average grayscale values ​​and the difference in the number of grayscale detection points, a first degree of proximity between grayscale detection point clusters is determined, and based on the relative position difference characteristics between the grayscale detection point clusters, the first degree of proximity is corrected to obtain a second degree of proximity, and if the second degree of proximity is greater than or equal to a preset value, the corresponding grayscale detection point clusters are associated;

[0022] Among them, the expression for calculating the second degree of proximity is:

[0023] ;

[0024] Among them, J is the second degree of closeness, is the weight coefficient affecting the difference in average gray value, is the weight coefficient affecting the difference in the number of grayscale detection points, is the preset maximum gray value difference, is the average gray value difference, is the preset difference in the number of grayscale detection points, is the difference in the number of grayscale detection points, is the relative position difference characteristic parameter between grayscale detection point clusters, F is the influence adjustment coefficient of the relative position difference characteristic parameter, and c is the influence adjustment constant of the relative position difference characteristic parameter.

[0025] In some embodiments disclosed in the present invention, the method for determining the relative position difference characteristics between grayscale detection point clusters includes:

[0026] Determine a central grayscale detection point of the grayscale detection point cluster, and radiate a number of distance detection lines from the central grayscale detection point to the edge grayscale detection points of the grayscale detection point cluster;

[0027] Based on the length of each distance detection line, determine the distance to shift the central gray detection point in the corresponding direction, and record the shifted central gray detection point as a representative gray detection point;

[0028] The relative distance of the representative grayscale detection points of the grayscale detection point cluster is calculated, and the relative distance is identified as a relative position difference characteristic parameter.

[0029] In some embodiments disclosed in the present invention, a method for constructing a dynamic representation of a grayscale representation extension model in a furnace includes:

[0030] Determine the mutually related grayscale detection point clusters, and connect the representative grayscale detection points corresponding to the mutually related grayscale detection point clusters to obtain cluster-related pointing lines, map the grayscale detection point clusters to the virtual smelting furnace three-dimensional model, and combine the cluster-related pointing lines to construct an extended model of the grayscale representation in the furnace;

[0031] A time reference axis is established, and based on the relative time nodes of shooting different infrared thermal imaging images, the corresponding furnace grayscale representation model is configured at the corresponding time nodes on the time reference axis, and the changes of the furnace grayscale representation model on the reference time axis are recognized as the dynamic representation of the furnace grayscale representation model.

[0032] In some embodiments disclosed in the present invention, the method for screening and classifying the dynamic performance of the extension model of the grayscale performance in the furnace includes:

[0033] The first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace includes:

[0034] The dynamic performance of the extended model of grayscale performance in the furnace is screened and classified by taking the cluster number interval to which the cluster number of the grayscale detection point cluster belongs as the screening condition, and the dynamic performance of the extended model of grayscale performance in the furnace is screened and classified by taking the detection point number interval to which the number of grayscale detection points in each grayscale detection point cluster belongs as the screening condition;

[0035] The method of secondary screening and classification of the dynamic performance of the extended model of grayscale performance in the furnace includes:

[0036] The dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification is analyzed, and the matching characteristics of the cluster-related pointing lines corresponding to each time node on the reference time axis are used as screening conditions to screen and classify the dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification;

[0037] The method for determining the matching features of cluster-associated pointing lines includes:

[0038] With the cluster-related pointing line as the axis line, a pointing column is constructed, and the cross volume between the pointing bodies corresponding to the cluster-related pointing lines participating in the comparison is calculated. If the ratio of the cross volume to the volume of the pointing column is greater than or equal to the preset value, the cluster-related pointing line is deemed to meet the initial matching conditions. If, during the dynamic performance comparison of the extended model of the grayscale performance in the furnace, the number of cluster-related pointing lines that meet the initial matching conditions is greater than or equal to the preset value, the extended model of the grayscale performance in the furnace to be compared is screened as one category.

[0039] The method of performing a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace classified by the second screening and classification includes:

[0040] A control time axis is constructed for the historical smelting furnace control parameter data, and different historical smelting control parameters are configured on the control time axis based on the time nodes corresponding to the historical smelting furnace control parameters at different time nodes;

[0041] Compare the historical smelting control parameters at different time nodes on the control time axis. If the difference between the historical smelting control parameters at the same time node is less than or equal to the preset value, the parameters are deemed to be close once. If the ratio of the number of times the parameters are deemed to be close to the number of all time nodes of the compared control time axis is greater than or equal to the preset value, the control parameters of the extended model of the grayscale performance in the furnace are deemed to be close. If the control parameters are close, the corresponding extended model of the grayscale performance in the furnace will be screened and classified.

[0042] In some embodiments disclosed in the present invention, the key factors of each virtual smelting furnace three-dimensional model dynamic performance set include:

[0043] The average gray value interval of the gray detection point group, the distribution ratio group interval of the gray detection point group, the gray detection point quantity interval of the gray detection point group, and the cluster correlation pointing lines between the gray detection point clusters.

[0044] In some embodiments disclosed in the present invention, a method for selecting a virtual smelting furnace three-dimensional model having the same dynamic performance as the extended model of the grayscale performance in the current furnace from the dynamic performance set of the virtual smelting furnace three-dimensional model comprises:

[0045] Compare the dynamic performance of the extended model of the grayscale performance in the furnace in the dynamic performance set of the virtual smelting furnace three-dimensional model with the dynamic performance of the extended model of the grayscale performance in the current furnace, determine the first difference parameter of the distribution ratio group of the grayscale detection point clusters at different time nodes, and determine the second difference parameter of the cluster-related pointing line;

[0046] Based on the first difference parameter and the second difference parameter, determine the degree of equivalence between the dynamic performance of the extended model of the grayscale performance in the furnace in the dynamic performance set of the virtual smelting furnace three-dimensional model and the dynamic performance of the extended model of the grayscale performance in the current furnace, and based on the degree of equivalence, determine the dynamic performance of the selected extended model of the grayscale performance in the furnace;

[0047] The method for determining the first difference parameter includes calculating the distribution ratio difference between each relative distribution ratio in the distribution ratio group, and determining the first difference parameter based on the distribution ratio difference corresponding to different distribution ratios;

[0048] The method for determining the second difference parameter includes calculating the matching parameters of the cluster-associated pointing lines corresponding to the extended model of the grayscale performance in the furnace at different time nodes, and the method for calculating the matching parameters includes taking the cluster-associated pointing lines as the axis line, constructing a pointing column, and calculating the cross volume between the pointing bodies corresponding to the cluster-associated pointing lines participating in the comparison, and calculating the cross volume and the volume ratio of the pointing column, identifying the volume ratio as the matching parameter, and determining the second difference parameter based on the matching parameters of all relevant pointing lines;

[0049] The expression for calculating the degree of equality is:

[0050] ;

[0051] Among them, W is the degree of equality, is the distribution ratio difference judgment function corresponding to the i-th distribution ratio in all distribution ratio groups. If the distribution ratio difference is less than or equal to the preset value, then Output 1, otherwise output 0, n is the total number of distribution ratios in all distribution ratio groups, is the volume ratio judgment function corresponding to the xth cluster-associated pointing line among all cluster-associated pointing lines. If the volume ratio is greater than or equal to the preset value, then Output 1, otherwise output 0, K is the volume ratio adjustment coefficient, and b is the volume ratio adjustment constant.

[0052] In some embodiments disclosed in the present invention, an industrial silicon smelting furnace control system based on infrared image analysis is also disclosed, including:

[0053] The first module is used to collect historical smelting furnace control parameter data, collect historical infrared thermal imaging image records at the same time, and collect simultaneous smelting records. Based on the analysis of the smelting records, the historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results are eliminated;

[0054] The second module is used to analyze the design structure image of the smelting furnace, and based on the analysis results, construct a virtual smelting furnace three-dimensional model, and based on the analysis of several historical infrared thermal imaging images at the same time, configure an infrared expression layer on the side of the virtual smelting furnace three-dimensional model, and grayscale the infrared expression layer to obtain a grayscale expression layer;

[0055] The third module is used to analyze the grayscale value characteristics of the grayscale expression layer, determine the grayscale value characteristics of different blocks, and perform correlation analysis on the grayscale value characteristics of the virtual smelting furnace three-dimensional model. Based on the analysis results, the furnace grayscale expression extension model is constructed, and the dynamic expression of the furnace grayscale expression extension model is constructed according to the temporal change of the furnace grayscale expression extension model.

[0056] The fourth module is used to perform a first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the grayscale value features of the grayscale performance layers, and perform a second screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the dynamic performances of the extended model of the grayscale performance in the furnace, and perform a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace that is screened and classified for the second time based on the equivalence between the historical smelting furnace control parameter data, to obtain a plurality of virtual smelting furnace three-dimensional model dynamic performance sets;

[0057] The fifth module is used to perform key factor analysis on the dynamic performance set of each virtual smelting furnace three-dimensional model. When the smelting furnace is controlled in the later stage, the matching between the key factors and the dynamic performance of the grayscale performance extension model in the current furnace constructed for the current smelting furnace is used to determine the virtual smelting furnace three-dimensional model dynamic performance set to be called, and the virtual smelting furnace three-dimensional model with the same dynamic performance as the grayscale performance extension model in the current furnace is screened out from the virtual smelting furnace three-dimensional model dynamic performance set, and the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model is used as a reference for the current control parameters of the smelting furnace.

[0058] The present invention discloses an industrial silicon smelting furnace control method and system based on infrared image analysis, which relates to the technical field of industrial silicon smelting furnace control. The smelting furnace design structure is analyzed, a virtual three-dimensional model is constructed, and an infrared representation layer is added to the model in combination with the infrared thermal imaging image, and then grayscale processing is performed; by analyzing the characteristics of the grayscale layer, a grayscale representation extension model in the furnace is constructed, and its dynamic changes are captured; based on the equality of grayscale values, dynamic representations and historical control parameter data, the model is screened and classified multiple times to form a number of virtual smelting furnace three-dimensional model dynamic representation sets; in the later smelting furnace control, key factors are used to match the current furnace conditions with the dynamic representation set, and virtual models with equivalent representations are screened out, and their historical control parameters are used as references to guide the real-time control of the smelting furnace. The present invention realizes accurate and efficient smelting process management through the above technical solutions.

[0059] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a method step diagram of the industrial silicon smelting furnace control method based on infrared image analysis disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0062] The following will be combined with the accompanying drawings and specific embodiments to clearly and completely describe the technical solution of the present invention. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention and cannot be understood as limiting the scope of protection of the present invention. Those skilled in the art in this field can make some non-essential improvements and adjustments based on the content of the present invention described below. In the present invention, unless otherwise clearly specified and limited, the technical terms used in the present invention should be the common meanings understood by the technical personnel described in the present invention.

[0063] Example:

[0064] The present invention discloses an industrial silicon smelting furnace control method based on infrared image analysis, see Figure 1 ,include:

[0065] Step S100, collecting historical smelting furnace control parameter data, collecting simultaneous historical infrared thermal imaging image records, and collecting simultaneous smelting records, based on the analysis of the smelting records, eliminating historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results.

[0066] First, through a high-precision data acquisition system, the control parameter data of the historical smelting furnace are comprehensively collected. These data include but are not limited to key indicators such as furnace temperature, furnace pressure, and feed rate. At the same time, infrared thermal imaging technology is used to capture the temperature distribution image in the furnace at the same time to form a historical infrared thermal imaging image record. In addition, the specific operations and results of each smelting are recorded in detail to form a smelting record. Subsequently, based on the analysis of the smelting records, the smelting results are strictly evaluated, and historical data with unqualified smelting results due to improper operation, equipment failure or raw material problems are eliminated. The purpose of this step is to ensure that the data relied on for subsequent model construction and analysis are highly accurate and representative, thereby providing a solid foundation for the precise control of the smelting furnace.

[0067] Step S200, analyze the design structure image of the smelting furnace, and based on the analysis results, construct a three-dimensional model of the virtual smelting furnace, and based on the analysis of several historical infrared thermal imaging images of the same moment, configure an infrared presentation layer on the side of the three-dimensional model of the virtual smelting furnace, and grayscale the infrared presentation layer to obtain a grayscale presentation layer.

[0068] Next, the design structure image of the smelting furnace is carefully analyzed, including the shape, size, material and layout of the internal components of the furnace. Based on these analysis results, a three-dimensional model of the virtual smelting furnace is constructed using three-dimensional modeling technology, which can truly reflect the physical structure of the smelting furnace. Then, combined with multiple historical infrared thermal imaging images at the same time, an infrared expression layer is configured on the side of the virtual smelting furnace three-dimensional model. This layer can intuitively display the distribution of the temperature field in the furnace. For the convenience of subsequent processing, the infrared expression layer is grayed and converted into a grayscale expression layer, which is convenient for image analysis and feature extraction.

[0069] In some embodiments disclosed in the present invention, a method for configuring an infrared display layer on the side of a virtual smelting furnace three-dimensional model includes:

[0070] Step S201, a positioning array is set for each historical infrared thermal imaging image, the positioning array includes a number of evenly arranged positioning points, and a corresponding positioning block is set for each positioning point, and each positioning point corresponds to a preset position on the side of the virtual smelting furnace three-dimensional model.

[0071] When configuring the infrared representation layer of the side of the virtual smelting furnace three-dimensional model, step S201 is first executed. The core of this step is to set an accurate positioning array for each historical infrared thermal imaging image. This positioning array consists of a number of evenly distributed positioning points, each of which corresponds to a specific area on the image, namely the positioning block. The design of these positioning points is very critical because they need to correspond one-to-one with the preset positions on the side of the virtual smelting furnace three-dimensional model to ensure that the infrared thermal imaging data can be accurately mapped to the model. In this way, it can be ensured that the spatial relationship between the infrared thermal imaging image and the three-dimensional model is accurately maintained.

[0072] Step S202 , intercepting the infrared thermal imaging block corresponding to the positioning block on the historical infrared thermal imaging image, and mapping the intercepted infrared thermal imaging block to the side of the virtual smelting furnace three-dimensional model.

[0073] In this step, according to the previously set positioning array, the infrared thermal imaging blocks corresponding to the positioning blocks are intercepted from the historical infrared thermal imaging images. These blocks contain key information about the temperature distribution in the furnace. Then, through advanced image processing technology, these intercepted infrared thermal imaging blocks are accurately mapped to the side of the virtual smelting furnace three-dimensional model. This mapping process needs to ensure that each infrared thermal imaging block is perfectly aligned with the corresponding position on the model to form a complete infrared performance layer. This layer not only intuitively shows the temperature distribution in the furnace, but also provides valuable data support for subsequent temperature field analysis and smelting furnace control. Through the close combination of these two steps, the effective fusion of infrared thermal imaging data and the virtual smelting furnace three-dimensional model can be achieved, laying a solid foundation for the precise control of the smelting process.

[0074] Step S300, performing grayscale value feature analysis on the grayscale representation layer, determining the grayscale value features of different blocks, and performing correlation analysis on the grayscale value features of the virtual smelting furnace three-dimensional model. Based on the analysis results, constructing a furnace grayscale representation extension model, and constructing a dynamic representation of the furnace grayscale representation extension model based on the temporal changes of the furnace grayscale representation extension model.

[0075] In some embodiments disclosed in the present invention, the method for performing grayscale value feature analysis on a grayscale representation layer includes:

[0076] Step S301, establish a number of circle analysis lines for the grayscale expression layer, and set a number of grayscale detection points symmetrically above and below the circle analysis lines, collect the grayscale value of each grayscale detection point, and calculate the grayscale value difference vector between adjacent grayscale detection points.

[0077] First, execute step S301 to establish several circle analysis lines on the grayscale representation layer. These analysis lines are like "rings" that divide the layer into different areas for local analysis. Then, several grayscale detection points are set at the upper and lower symmetrical positions of each analysis line. These points are the key positions for collecting grayscale values. By collecting the grayscale value of each grayscale detection point and calculating the grayscale value difference vector between adjacent points, we can preliminarily understand the distribution of grayscale values ​​on the layer and its changing trend.

[0078] Step S302 , performing continuity analysis on the grayscale value difference vectors, if the change directions of adjacent grayscale value difference vectors are the same and the values ​​of the grayscale value difference vectors are less than or equal to a preset value, the corresponding grayscale detection points are associated until a grayscale detection point cluster is formed.

[0079] Go to step S302, and perform continuity analysis on these grayscale value difference vectors. If the change directions of adjacent grayscale value difference vectors are the same, and their values ​​are less than or equal to the preset threshold, then it is considered that there is a certain correlation between these grayscale detection points, and they may belong to the same temperature area or the same thermal effect area. Based on this correlation, these points are connected to form a grayscale detection point cluster. This helps to more clearly identify different temperature areas or thermal effect patterns on the layer.

[0080] Step S303, calculating the distribution ratio of each grayscale detection point cluster above and below different circle analysis lines to obtain a distribution ratio group, and associating the distribution ratio group with the grayscale detection point cluster.

[0081] Execute step S303 to calculate the distribution ratio of each grayscale detection point cluster on different circle analysis lines. This ratio can reflect the spatial position characteristics of the cluster on the layer. This ratio is associated with the grayscale detection point cluster to provide more information for subsequent feature analysis.

[0082] Step S304, analyzing the number of grayscale detection points contained in each grayscale detection point cluster, and calculating the average grayscale value of the grayscale detection point cluster, and identifying the distribution ratio group, the number of grayscale detection points and the average grayscale value as grayscale value features.

[0083] Enter step S304, count the number of grayscale detection points in each grayscale detection point cluster, and calculate the average grayscale value of the cluster. These two indicators can reflect the scale and overall grayscale level of the cluster. By combining the distribution ratio group, the number of grayscale detection points and the average grayscale value, the grayscale value characteristics of the grayscale representation layer are obtained. These characteristics provide strong data support for subsequent image analysis, temperature field simulation and smelting furnace control.

[0084] In some embodiments disclosed in the present invention, a method for performing correlation analysis on grayscale value features of a virtual smelting furnace three-dimensional model includes:

[0085] Step S305, based on the difference in average grayscale values ​​and the difference in the number of grayscale detection points, determine the first degree of proximity between the grayscale detection point clusters, and based on the relative position difference characteristics between the grayscale detection point clusters, correct the first degree of proximity to obtain a second degree of proximity. If the second degree of proximity is greater than or equal to a preset value, the corresponding grayscale detection point clusters are associated.

[0086] Among them, the expression for calculating the second degree of proximity is:

[0087] .

[0088] Among them, J is the second degree of closeness, is the weight coefficient affecting the difference in average gray value, is the weight coefficient affecting the difference in the number of grayscale detection points, is the preset maximum gray value difference, is the average gray value difference, is the preset difference in the number of grayscale detection points, is the difference in the number of grayscale detection points, is the relative position difference characteristic parameter between grayscale detection point clusters, F is the influence adjustment coefficient of the relative position difference characteristic parameter, and c is the influence adjustment constant of the relative position difference characteristic parameter.

[0089] In some embodiments disclosed in the present invention, the method for determining the relative position difference characteristics between grayscale detection point clusters includes:

[0090] Step S3051, determine the central grayscale detection point of the grayscale detection point cluster, and radiate a number of distance detection lines from the central grayscale detection point to the edge grayscale detection points of the grayscale detection point cluster.

[0091] Step S3052: Based on the length of each distance detection line, determine the distance to shift the central grayscale detection point in the corresponding direction, and record the shifted central grayscale detection point as the representative grayscale detection point.

[0092] Step S3052, calculating the relative distance of the representative grayscale detection points of the grayscale detection point cluster, and identifying the relative distance as a relative position difference feature parameter.

[0093] In some embodiments disclosed in the present invention, a method for constructing a dynamic representation of a grayscale representation extension model in a furnace includes:

[0094] Step S306, determine the mutually related grayscale detection point clusters, and connect the representative grayscale detection points corresponding to the mutually related grayscale detection point clusters to obtain cluster-related pointing lines, map the grayscale detection point clusters to the virtual smelting furnace three-dimensional model, and combine the cluster-related pointing lines to construct an extended model of the grayscale representation in the furnace.

[0095] Execute step S306 to determine those interrelated grayscale detection point clusters. These clusters show similar grayscale value characteristics on the grayscale representation layer, indicating that they may belong to the same temperature zone or be affected by the same thermal effect. In order to more intuitively display the relationship between these clusters, the representative grayscale detection points of each cluster (usually the center point of the cluster or a representative point) are connected to form cluster-associated pointing lines. These pointing lines not only reveal the spatial relationship between clusters, but also provide a basis for the subsequent mapping of grayscale information to the three-dimensional model of the virtual smelting furnace. Next, these grayscale detection point clusters and their associated pointing lines are accurately mapped to the three-dimensional model of the virtual smelting furnace, thereby constructing an extended model of grayscale representation in the furnace. This model not only contains the grayscale distribution information in the furnace, but also shows the intrinsic connection between grayscale values ​​through associated pointing lines.

[0096] Step S307, establish a time reference axis, and based on the relative time nodes of shooting different infrared thermal imaging images, configure the corresponding furnace grayscale representation model at the corresponding time nodes on the time reference axis, and recognize the changes of the furnace grayscale representation model on the reference time axis as the dynamic representation of the furnace grayscale representation model.

[0097] Going to step S307, in order to capture the changes in the grayscale performance in the furnace over time, a time reference axis is established. This axis is like a time ruler, which helps to record and analyze the grayscale performance at different times. According to the shooting time node of each infrared thermal imaging image, the corresponding furnace grayscale performance model is configured at the corresponding time node on the time reference axis. In this way, it is possible to clearly see the changes in the furnace grayscale performance model over time, that is, its dynamic performance. By analyzing these dynamic performances, we can have a deeper understanding of the evolution of the temperature field in the furnace during the smelting process, and provide strong data support for the precise control of the smelting furnace.

[0098] Step S400, based on the equivalence between the grayscale value features of the grayscale representation layers, the dynamic representation of the extended model of the grayscale representation in the furnace is screened and classified for the first time, and based on the equivalence between the dynamic representations of the extended model of the grayscale representation in the furnace, the dynamic representation of the extended model of the grayscale representation in the furnace is screened and classified for the second time, and based on the equivalence between the control parameter data of the historical smelting furnaces, the dynamic representation of the extended model of the grayscale representation in the furnace that is screened and classified for the second time is screened and classified for the third time, so as to obtain a number of dynamic representation sets of virtual smelting furnace three-dimensional models.

[0099] Based on the dynamic performance set of the extended model of the grayscale performance in the furnace, multiple screening and classification are performed. First, based on the equivalence between the grayscale value features, the dynamic performance is screened and classified for the first time to eliminate abnormal data that does not conform to the mainstream characteristics. Then, based on the equivalence between the dynamic performances, a second screening and classification is performed to classify the performances with similar dynamic change characteristics into one category. Finally, a third screening and classification is performed in combination with the equivalence between the historical smelting furnace control parameter data to ensure that each type of dynamic performance set corresponds to a specific smelting condition and control parameter combination. Through this step, several representative and reusable dynamic performance sets of virtual smelting furnace three-dimensional models can be obtained.

[0100] In some embodiments disclosed in the present invention, the method for screening and classifying the dynamic performance of the extension model of the grayscale performance in the furnace includes:

[0101] The first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace includes:

[0102] Step S401, using the cluster number interval to which the cluster number of the grayscale detection point cluster belongs as a screening condition, to screen and classify the dynamic performance of the extended model of the grayscale performance in the furnace, and using the detection point number interval to which the number of grayscale detection points in each grayscale detection point cluster belongs as a screening condition, to screen and classify the dynamic performance of the extended model of the grayscale performance in the furnace.

[0103] The method of secondary screening and classification of the dynamic performance of the extended model of grayscale performance in the furnace includes:

[0104] Step S402, analyzing the dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification, and screening and classifying the dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification based on the matching characteristics of the cluster-associated pointing lines corresponding to each time node on the reference time axis.

[0105] The method for determining the matching features of cluster-associated pointing lines includes:

[0106] With the cluster-associated pointing lines as the axis line, a pointing column is constructed, and the intersection volume between the pointing bodies corresponding to the cluster-associated pointing lines participating in the comparison is calculated. If the ratio of the intersection volume to the volume of the pointing column is greater than or equal to the preset value, the cluster-associated pointing lines are deemed to meet the initial matching conditions. If, during the dynamic performance comparison of the in-furnace grayscale performance extension model, the number of cluster-associated pointing lines that meet the initial matching conditions is greater than or equal to the preset value, the compared in-furnace grayscale performance extension model will be screened as one category.

[0107] The method of performing a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace classified by the second screening and classification includes:

[0108] Step S403, constructing a control time axis for the historical smelting furnace control parameter data, and configuring different historical smelting control parameters on the control time axis based on the time nodes corresponding to the historical smelting furnace control parameters at different time nodes.

[0109] Step S404, compare the historical smelting control parameters of different time nodes on the control time axis. If the difference between the historical smelting control parameters at the same time node is less than or equal to the preset value, the parameters are deemed to be close once. If the ratio of the number of times the parameters are deemed to be close to the number of all time nodes of the compared control time axis is greater than or equal to the preset value, the control parameters of the extended model of the grayscale performance in the furnace are deemed to be close. If the control parameters are close, the corresponding extended model of the grayscale performance in the furnace will be screened and classified.

[0110] Step S500, perform key factor analysis on each virtual smelting furnace three-dimensional model dynamic performance set, and when controlling the smelting furnace in the later stage, use the matching of the key factors with the dynamic performance of the current furnace grayscale performance extension model constructed for the current smelting furnace to determine the called virtual smelting furnace three-dimensional model dynamic performance set, and screen out the virtual smelting furnace three-dimensional model with the same dynamic performance as the current furnace grayscale performance extension model from the virtual smelting furnace three-dimensional model dynamic performance set, and use the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model as a reference for the current control parameters of the smelting furnace.

[0111] When controlling the smelting furnace in the later stage, the key factor analysis technology is first used to determine the key factors and their weights that affect the smelting results. Then, the furnace condition information of the current smelting furnace is matched with the dynamic performance set of the virtual smelting furnace three-dimensional model to find the dynamic performance set that is most similar to the current furnace condition. In this performance set, the virtual smelting furnace three-dimensional model that is equivalent to the dynamic performance of the grayscale performance extension model in the current furnace is further screened out. Finally, with the historical smelting control parameter data corresponding to the model as a reference, appropriate adjustments and optimizations are made in combination with the actual situation of the current smelting furnace, so as to achieve accurate and efficient control of the smelting furnace. This step makes full use of the advantages of historical data and model predictions, and improves the stability and controllability of the smelting process.

[0112] In some embodiments disclosed in the present invention, a method for selecting a virtual smelting furnace three-dimensional model having the same dynamic performance as the extended model of the grayscale performance in the current furnace from the dynamic performance set of the virtual smelting furnace three-dimensional model comprises:

[0113] Step S501, compare the dynamic performance of the extended model of the grayscale performance in the furnace in the dynamic performance set of the virtual smelting furnace three-dimensional model with the dynamic performance of the extended model of the grayscale performance in the current furnace, determine the first difference parameter of the distribution ratio group of the grayscale detection point clusters at different time nodes, and determine the second difference parameter of the cluster-related pointing lines.

[0114] Step S502, based on the first difference parameter and the second difference parameter, determine the degree of equivalence between the dynamic performance of the in-furnace grayscale performance extension model in the dynamic performance set of the virtual smelting furnace three-dimensional model and the dynamic performance of the current in-furnace grayscale performance extension model, and based on the degree of equivalence, determine the dynamic performance of the selected in-furnace grayscale performance extension model.

[0115] The method for determining the first difference parameter includes calculating the distribution ratio difference between each relative distribution ratio in the distribution ratio group, and determining the first difference parameter based on the distribution ratio difference corresponding to different distribution ratios.

[0116] Among them, the method for determining the second difference parameter includes calculating the matching parameters of the cluster-associated pointing lines corresponding to the extension model of the grayscale performance in the furnace at different time nodes, and the method for calculating the matching parameters includes taking the cluster-associated pointing lines as the axis line, constructing a pointing column, and calculating the intersection volume between the pointing bodies corresponding to the cluster-associated pointing lines participating in the comparison, and calculating the ratio of the intersection volume to the volume of the pointing column, identifying the volume ratio as the matching parameter, and determining the second difference parameter based on the matching parameters of all relevant pointing lines.

[0117] The expression for calculating the degree of equality is:

[0118] .

[0119] Among them, W is the degree of equality, is the distribution ratio difference judgment function corresponding to the i-th distribution ratio in all distribution ratio groups. If the distribution ratio difference is less than or equal to the preset value, then Output 1, otherwise output 0, n is the total number of distribution ratios in all distribution ratio groups, is the volume ratio judgment function corresponding to the xth cluster-associated pointing line among all cluster-associated pointing lines. If the volume ratio is greater than or equal to the preset value, then Output 1, otherwise output 0, K is the volume ratio adjustment coefficient, and b is the volume ratio adjustment constant.

[0120] In some embodiments disclosed in the present invention, the key factors of each virtual smelting furnace three-dimensional model dynamic performance set include:

[0121] The average gray value interval of the gray detection point group, the distribution ratio group interval of the gray detection point group, the gray detection point quantity interval of the gray detection point group, and the cluster correlation pointing lines between the gray detection point clusters.

[0122] In some embodiments disclosed in the present invention, an industrial silicon smelting furnace control system based on infrared image analysis is also disclosed, including:

[0123] The first module is used to collect historical smelting furnace control parameter data, collect historical infrared thermal imaging image records at the same time, and collect simultaneous smelting records. Based on the analysis of the smelting records, the historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results are eliminated;

[0124] The second module is used to analyze the design structure image of the smelting furnace, and based on the analysis results, construct a virtual smelting furnace three-dimensional model, and based on the analysis of several historical infrared thermal imaging images at the same time, configure an infrared expression layer on the side of the virtual smelting furnace three-dimensional model, and grayscale the infrared expression layer to obtain a grayscale expression layer;

[0125] The third module is used to analyze the grayscale value characteristics of the grayscale expression layer, determine the grayscale value characteristics of different blocks, and perform correlation analysis on the grayscale value characteristics of the virtual smelting furnace three-dimensional model. Based on the analysis results, the furnace grayscale expression extension model is constructed, and the dynamic expression of the furnace grayscale expression extension model is constructed according to the temporal change of the furnace grayscale expression extension model.

[0126] The fourth module is used to perform a first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the grayscale value features of the grayscale performance layers, and perform a second screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the dynamic performances of the extended model of the grayscale performance in the furnace, and perform a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace that is screened and classified for the second time based on the equivalence between the historical smelting furnace control parameter data, to obtain a plurality of virtual smelting furnace three-dimensional model dynamic performance sets;

[0127] The fifth module is used to perform key factor analysis on the dynamic performance set of each virtual smelting furnace three-dimensional model. When the smelting furnace is controlled in the later stage, the matching between the key factors and the dynamic performance of the grayscale performance extension model in the current furnace constructed for the current smelting furnace is used to determine the virtual smelting furnace three-dimensional model dynamic performance set to be called, and the virtual smelting furnace three-dimensional model with the same dynamic performance as the grayscale performance extension model in the current furnace is screened out from the virtual smelting furnace three-dimensional model dynamic performance set, and the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model is used as a reference for the current control parameters of the smelting furnace.

[0128] The present invention discloses an industrial silicon smelting furnace control method and system based on infrared image analysis, which relates to the technical field of industrial silicon smelting furnace control. The smelting furnace design structure is analyzed, a virtual three-dimensional model is constructed, and an infrared representation layer is added to the model in combination with the infrared thermal imaging image, and then grayscale processing is performed; by analyzing the characteristics of the grayscale layer, a grayscale representation extension model in the furnace is constructed, and its dynamic changes are captured; based on the equality of grayscale values, dynamic representations and historical control parameter data, the model is screened and classified multiple times to form a number of virtual smelting furnace three-dimensional model dynamic representation sets; in the later smelting furnace control, key factors are used to match the current furnace conditions with the dynamic representation set, and virtual models with equivalent representations are screened out, and their historical control parameters are used as references to guide the real-time control of the smelting furnace. The present invention realizes accurate and efficient smelting process management through the above technical solutions.

[0129] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or by software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. An industrial silicon smelting furnace control method based on infrared image analysis, characterized in that: include: Collect historical smelting furnace control parameter data, collect historical infrared thermal imaging image records at the same time, and collect simultaneous smelting records. Based on the analysis of the smelting records, eliminate the historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results; The design structure image of the smelting furnace is analyzed, and based on the analysis result, a virtual smelting furnace three-dimensional model is constructed, and based on the analysis of several historical infrared thermal imaging images at the same time, an infrared expression layer is configured on the side of the virtual smelting furnace three-dimensional model, and the infrared expression layer is grayed to obtain a gray expression layer; The grayscale value characteristics of the grayscale expression layer are analyzed to determine the grayscale value characteristics of different blocks, and the grayscale value characteristics of the virtual smelting furnace three-dimensional model are analyzed for correlation. Based on the analysis results, the furnace grayscale expression extension model is constructed, and the dynamic expression of the furnace grayscale expression extension model is constructed according to the temporal change of the furnace grayscale expression extension model. Based on the equivalence between the grayscale value features of the grayscale representation layers, the dynamic representation of the grayscale representation extension model in the furnace is first screened and classified, and based on the equivalence between the dynamic representations of the grayscale representation extension model in the furnace, the dynamic representation of the grayscale representation extension model in the furnace is secondly screened and classified, and based on the equivalence between the historical smelting furnace control parameter data, the dynamic representation of the grayscale representation extension model in the furnace that is screened and classified for the second time is thirdly screened and classified to obtain a number of virtual smelting furnace three-dimensional model dynamic representation sets; A key factor analysis is performed on the dynamic performance set of each virtual smelting furnace three-dimensional model. When the smelting furnace is controlled in the later stage, the matching between the key factors and the dynamic performance of the grayscale performance extension model in the current furnace constructed for the current smelting furnace is used to determine the dynamic performance set of the virtual smelting furnace three-dimensional model to be called, and the virtual smelting furnace three-dimensional model with the same dynamic performance as the grayscale performance extension model in the current furnace is screened out from the dynamic performance set of the virtual smelting furnace three-dimensional model, and the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model provides a reference for the current control parameters of the smelting furnace.

2. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 1 is characterized in that: The method for configuring an infrared representation layer on the side of a virtual smelting furnace three-dimensional model includes: A positioning array is set for each historical infrared thermal imaging image, the positioning array includes a number of evenly arranged positioning points, and a corresponding positioning block is set for each positioning point, and each positioning point corresponds to a preset position on the side of the virtual smelting furnace three-dimensional model; The infrared thermal imaging block corresponding to the positioning block on the historical infrared thermal imaging image is intercepted, and the intercepted infrared thermal imaging block is mapped to the side of the virtual smelting furnace three-dimensional model.

3. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 1 is characterized in that: Methods for analyzing grayscale value characteristics of grayscale representation layers include: Establishing a number of circle analysis lines for the grayscale expression layer, and setting a number of grayscale detection points symmetrically above and below the circle analysis line, collecting the grayscale value of each grayscale detection point, and calculating the grayscale value difference vector between adjacent grayscale detection points; Perform continuity analysis on the grayscale value difference vectors. If the change directions of adjacent grayscale value difference vectors are the same and the values ​​of the grayscale value difference vectors are less than or equal to a preset value, then associate the corresponding grayscale detection points until a grayscale detection point cluster is formed. Calculate the distribution ratio of each grayscale detection point cluster above and below different circle analysis lines to obtain a distribution ratio group, and associate the distribution ratio group with the grayscale detection point cluster; The number of grayscale detection points contained in each grayscale detection point cluster is analyzed, and the average grayscale value of the grayscale detection point cluster is calculated, and the distribution ratio group, the number of grayscale detection points and the average grayscale value are identified as grayscale value features.

4. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 3 is characterized in that: The method for performing correlation analysis on the gray value characteristics of the virtual smelting furnace three-dimensional model includes: Based on the difference in average grayscale values ​​and the difference in the number of grayscale detection points, a first degree of proximity between grayscale detection point clusters is determined, and based on the relative position difference characteristics between the grayscale detection point clusters, the first degree of proximity is corrected to obtain a second degree of proximity, and if the second degree of proximity is greater than or equal to a preset value, the corresponding grayscale detection point clusters are associated; Among them, the expression for calculating the second degree of proximity is: ； Among them, J is the second degree of closeness, is the weight coefficient affecting the difference in average gray value, is the weight coefficient affecting the difference in the number of grayscale detection points, is the preset maximum gray value difference, is the average gray value difference, is the preset difference in the number of grayscale detection points, is the difference in the number of grayscale detection points, is the relative position difference characteristic parameter between grayscale detection point clusters, F is the influence adjustment coefficient of the relative position difference characteristic parameter, and c is the influence adjustment constant of the relative position difference characteristic parameter.

5. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 4 is characterized in that: Methods for determining relative position difference characteristics between grayscale detection point clusters include: Determine a central grayscale detection point of the grayscale detection point cluster, and radiate a number of distance detection lines from the central grayscale detection point to the edge grayscale detection points of the grayscale detection point cluster; Based on the length of each distance detection line, determine the distance to shift the central gray detection point in the corresponding direction, and record the shifted central gray detection point as a representative gray detection point; The relative distance of the representative grayscale detection points of the grayscale detection point cluster is calculated, and the relative distance is identified as a relative position difference characteristic parameter.

6. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 5 is characterized in that: The method of constructing the dynamic representation of the extended model of the grayscale representation in the furnace includes: Determine the mutually related grayscale detection point clusters, and connect the representative grayscale detection points corresponding to the mutually related grayscale detection point clusters to obtain cluster-related pointing lines, map the grayscale detection point clusters to the virtual smelting furnace three-dimensional model, and combine the cluster-related pointing lines to construct an extended model of the grayscale representation in the furnace; A time reference axis is established, and based on the relative time nodes of shooting different infrared thermal imaging images, the corresponding furnace grayscale representation model is configured at the corresponding time nodes on the time reference axis, and the changes of the furnace grayscale representation model on the reference time axis are recognized as the dynamic representation of the furnace grayscale representation model.

7. The method for controlling an industrial silicon smelting furnace based on infrared image analysis according to claim 6, characterized in that: Methods for screening and classifying the dynamic performance of the extended model of grayscale performance in the furnace include: The first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace includes: The dynamic performance of the extended model of grayscale performance in the furnace is screened and classified by taking the cluster number interval to which the cluster number of the grayscale detection point cluster belongs as the screening condition, and the dynamic performance of the extended model of grayscale performance in the furnace is screened and classified by taking the detection point number interval to which the number of grayscale detection points in each grayscale detection point cluster belongs as the screening condition; The method of secondary screening and classification of the dynamic performance of the extended model of grayscale performance in the furnace includes: The dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification is analyzed, and the matching characteristics of the cluster-related pointing lines corresponding to each time node on the reference time axis are used as screening conditions to screen and classify the dynamic performance of the extended model of the grayscale performance in the furnace after the first screening and classification; The method for determining the matching features of cluster-associated pointing lines includes: With the cluster-related pointing lines as the axis, a pointing column is constructed, and the cross volume between the pointing bodies corresponding to the cluster-related pointing lines participating in the comparison is calculated. If the cross volume accounts for a volume ratio of the pointing column that is greater than or equal to a preset value, the cluster-related pointing lines are deemed to meet the initial matching conditions. If, during the comparison of the dynamic performance of the extended model of the grayscale performance in the furnace, the number of cluster-related pointing lines that meet the initial matching conditions is greater than or equal to a preset value, the extended model of the grayscale performance in the furnace that is compared is selected as one category. The method of performing a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace classified by the second screening and classification includes: A control time axis is constructed for the historical smelting furnace control parameter data, and different historical smelting control parameters are configured on the control time axis based on the time nodes corresponding to the historical smelting furnace control parameters at different time nodes; Compare the historical smelting control parameters at different time nodes on the control time axis. If the difference between the historical smelting control parameters at the same time node is less than or equal to the preset value, the parameters are deemed to be close once. If the ratio of the number of times the parameters are deemed to be close to the number of all time nodes of the compared control time axis is greater than or equal to the preset value, the control parameters of the extended model of the grayscale performance in the furnace are deemed to be close. If the control parameters are close, the corresponding extended model of the grayscale performance in the furnace will be screened and classified.

8. The method for controlling an industrial silicon smelting furnace based on infrared image analysis according to claim 6, characterized in that: The key factors of each virtual smelting furnace 3D model dynamic performance set include: The average gray value interval of the gray detection point group, the distribution ratio group interval of the gray detection point group, the gray detection point quantity interval of the gray detection point group, and the cluster correlation pointing lines between the gray detection point clusters.

9. The industrial silicon smelting furnace control method based on infrared image analysis according to claim 6 is characterized in that: The method of selecting a virtual smelting furnace three-dimensional model having the same dynamic performance as the extended model of the grayscale performance in the current furnace from the dynamic performance set of the virtual smelting furnace three-dimensional model includes: Compare the dynamic performance of the extended model of the grayscale performance in the furnace in the dynamic performance set of the virtual smelting furnace three-dimensional model with the dynamic performance of the extended model of the grayscale performance in the current furnace, determine the first difference parameter of the distribution ratio group of the grayscale detection point clusters at different time nodes, and determine the second difference parameter of the cluster-related pointing line; Based on the first difference parameter and the second difference parameter, determine the degree of equivalence between the dynamic performance of the extended model of the grayscale performance in the furnace in the dynamic performance set of the virtual smelting furnace three-dimensional model and the dynamic performance of the extended model of the grayscale performance in the current furnace, and based on the degree of equivalence, determine the dynamic performance of the selected extended model of the grayscale performance in the furnace; The method for determining the first difference parameter includes calculating the distribution ratio difference between each relative distribution ratio in the distribution ratio group, and determining the first difference parameter based on the distribution ratio difference corresponding to different distribution ratios; The method for determining the second difference parameter includes calculating the matching parameters of the cluster-associated pointing lines corresponding to the extended model of the grayscale performance in the furnace at different time nodes, and the method for calculating the matching parameters includes taking the cluster-associated pointing lines as the axis line, constructing a pointing column, and calculating the cross volume between the pointing bodies corresponding to the cluster-associated pointing lines participating in the comparison, and calculating the cross volume and the volume ratio of the pointing column, identifying the volume ratio as the matching parameter, and determining the second difference parameter based on the matching parameters of all relevant pointing lines; The expression for calculating the degree of equality is: ； Among them, W is the degree of equality, is the distribution ratio difference judgment function corresponding to the i-th distribution ratio in all distribution ratio groups. If the distribution ratio difference is less than or equal to the preset value, then Output 1, otherwise output 0, n is the total number of distribution ratios in all distribution ratio groups, is the volume ratio judgment function corresponding to the xth cluster-associated pointing line among all cluster-associated pointing lines. If the volume ratio is greater than or equal to the preset value, then Output 1, otherwise output 0, K is the volume ratio adjustment coefficient, and b is the volume ratio adjustment constant.

10. Industrial silicon smelting furnace control system based on infrared image analysis, characterized in that: include: The first module is used to collect historical smelting furnace control parameter data, collect historical infrared thermal imaging image records at the same time, and collect simultaneous smelting records. Based on the analysis of the smelting records, the historical smelting control parameter data and historical infrared thermal imaging image records with unqualified smelting results are eliminated; The second module is used to analyze the design structure image of the smelting furnace, and based on the analysis results, construct a virtual smelting furnace three-dimensional model, and based on the analysis of several historical infrared thermal imaging images at the same time, configure an infrared expression layer on the side of the virtual smelting furnace three-dimensional model, and grayscale the infrared expression layer to obtain a grayscale expression layer; The third module is used to analyze the grayscale value characteristics of the grayscale expression layer, determine the grayscale value characteristics of different blocks, and perform correlation analysis on the grayscale value characteristics of the virtual smelting furnace three-dimensional model. Based on the analysis results, the furnace grayscale expression extension model is constructed, and the dynamic expression of the furnace grayscale expression extension model is constructed according to the temporal change of the furnace grayscale expression extension model. The fourth module is used to perform a first screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the grayscale value features of the grayscale performance layers, and perform a second screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace based on the equivalence between the dynamic performances of the extended model of the grayscale performance in the furnace, and perform a third screening and classification of the dynamic performance of the extended model of the grayscale performance in the furnace that is screened and classified for the second time based on the equivalence between the historical smelting furnace control parameter data, to obtain a plurality of virtual smelting furnace three-dimensional model dynamic performance sets; The fifth module is used to perform key factor analysis on the dynamic performance set of each virtual smelting furnace three-dimensional model. When the smelting furnace is controlled in the later stage, the matching between the key factors and the dynamic performance of the grayscale performance extension model in the current furnace constructed for the current smelting furnace is used to determine the virtual smelting furnace three-dimensional model dynamic performance set to be called, and the virtual smelting furnace three-dimensional model with the same dynamic performance as the grayscale performance extension model in the current furnace is screened out from the virtual smelting furnace three-dimensional model dynamic performance set, and the historical smelting control parameter data corresponding to the virtual smelting furnace three-dimensional model is used as a reference for the current control parameters of the smelting furnace.

Images