Graphite composite insulation board production process control method, control system and storage medium

Abstract

The present application relates to the field of image processing technology, and discloses a graphite composite insulation board production process control method, control system and storage medium. The control method includes: obtaining pretreated raw materials by performing feature analysis and ratio optimization on the raw materials. Monitoring the reaction state of the pretreated raw materials, adjusting the process parameters, and generating modified raw materials. Formula optimization and performance prediction are performed based on the material parameters of the modified raw materials to obtain optimized formula materials. Analyze and optimize the molding process parameters of the formula materials, adjust the molding process, and generate molding materials. Optimize the curing process using the state parameters of the molding materials to form a cured product. Finally, evaluate the quality based on the test data of the cured product and generate a production control plan. The present application realizes intelligent control of the entire production process by establishing a multi-dimensional data acquisition and processing system, thereby improving the stability and consistency of product quality.

Description

Technical Field

[0001] The present application relates to the field of image processing, and in particular to a graphite composite insulation board production process control method, control system and storage medium. Background Art

[0002] Graphite composite insulation board is widely used in building energy conservation, industrial insulation and other fields. The existing production process of graphite composite insulation board is mainly carried out by manual experience control and single parameter monitoring. In the production process, the raw materials are processed by setting fixed process parameters such as temperature, pressure and time. Quality control mainly relies on manual experience judgment and offline testing, including the detection of indicators such as product appearance, size and thermal conductivity. The adjustment of production process parameters is mainly based on empirical data and historical production records, lacking real-time data support and scientific optimization methods.

[0003] This traditional production control method has obvious shortcomings. First, the manual experience control method is difficult to adapt to the needs of large-scale industrial production, with low production efficiency and large quality fluctuations. Secondly, the monitoring methods of the physical parameters and internal structural characteristics of materials in the prior art are relatively single, which cannot fully reflect the state changes of materials in the production process, and it is difficult to achieve precise control of the entire production process, resulting in unstable product quality. Thirdly, the lack of in-depth analysis of the correlation between multiple process parameters makes it difficult to achieve optimal regulation of process parameters. In addition, traditional quality control methods mainly rely on final inspection of products, and cannot promptly detect and deal with abnormal situations in the production process, resulting in a high scrap rate. Due to the lack of real-time data collection and analysis methods, it is difficult for the prior art to establish a complete quality traceability system, which affects the continuous improvement of product quality. Summary of the invention

[0004] The present application provides a graphite composite insulation board production process control method, control system and storage medium, which realizes intelligent control of the entire production process by establishing a multi-dimensional data acquisition and processing system, thereby improving the stability and consistency of product quality. This method realizes precise control of process parameters and real-time monitoring of quality through multi-field coupling calculation and data fusion analysis, solving the problems of high dependence on manual experience, insufficient parameter optimization, and lagging quality control in traditional production methods.

[0005] In a first aspect, the present application provides a method for controlling a production process of a graphite composite insulation board, comprising the following steps:

[0006] S1, raw material analysis stage: analyzing and processing the raw materials to obtain raw material characteristic data, optimizing the proportioning parameters based on the raw material characteristic data to obtain pre-treated raw materials;

[0007] S2, monitoring and processing stage: monitoring and processing are performed according to the state parameters corresponding to the pre-treated raw materials to obtain reaction state data, and the process parameters are adjusted and processed using the reaction state data to form modified raw materials;

[0008] S3, formulation optimization stage: performing formulation optimization processing according to the material parameters of the modified raw material to obtain formulation data, completing material performance prediction processing according to the formulation data, and generating optimized formulation materials;

[0009] S4, molding analysis stage: performing molding analysis processing through the physical parameters of the optimized formula material to obtain molding process parameters, and implementing process control processing based on the molding process parameters to produce molding materials;

[0010] S5, curing analysis stage: using the state parameters of the molding material to perform curing analysis processing to obtain curing process parameters, and implementing curing process control processing according to the curing process parameters to form a cured product;

[0011] S6, quality evaluation stage: performing quality evaluation processing based on the inspection data of the cured product to obtain evaluation parameters, executing control strategy generation processing according to the evaluation parameters to generate a production control plan.

[0012] The raw material analysis stage of step S1 includes the following steps:

[0013] S101, collecting spectra of the raw material in multiple time periods, and obtaining spectral characteristic data of the raw material by comparing the spectral intensity with the standard spectrum;

[0014] S102, calculating the time distribution of the physicochemical characteristics, particle size distribution and morphological characteristics according to the raw material spectral characteristic data, and generating initial characteristic standard data;

[0015] S103, performing partition calculation on the initial characteristic standard data according to 64 data channels to obtain partition characteristic data;

[0016] S104, performing cumulative calculation on the physical features, chemical features and morphological features in the partition feature data to form a feature cumulative spectrum;

[0017] S105, using the characteristic cumulative spectrum to perform distribution calculation on each data channel of the raw material to obtain raw material characteristic data;

[0018] S106, performing correlation analysis on the raw material characteristic data and the ratio optimization parameters, and outputting the pre-processed raw material.

[0019] The monitoring and processing stage of step S2 includes the following steps:

[0020] S201. The pretreated raw material is subjected to temperature field acquisition in multiple time periods, and temperature change data is obtained by comparative analysis of the temperature field intensity and the standard temperature field;

[0021] S202. Calculate the time distribution characteristics of the reaction rate, conversion rate and dispersion degree according to the temperature change data to generate reaction state data;

[0022] S203. The reaction state data is divided into regions according to 256 data sampling points to obtain partitioned reaction state data;

[0023] S204. Perform weighted cumulative calculation on the reaction rate, conversion rate and dispersion degree in the partition reaction state data to form a state cumulative distribution;

[0024] S205. Using the state cumulative distribution to perform distribution fitting calculation on the state parameters of each sampling point to obtain reaction state data;

[0025] S206. Compare and analyze the reaction state data with the process parameter threshold range, and output the modified raw material.

[0026] The recipe optimization stage of step S3 includes the following steps:

[0027] S301. Collecting the physical property parameters of the modified raw material in multiple time periods, and obtaining a material parameter sequence by comparative analysis of thermal conductivity, density and strength;

[0028] S302. Calculate the time variation characteristics of material strength, thermal conductivity and density distribution according to the material parameter sequence to generate formulation data;

[0029] S303. The formula data is numerically decomposed and calculated according to 512 sampling points to obtain partition formula data;

[0030] S304. Perform weighted statistical calculations on the physical properties, mechanical properties and thermal conductivity in the partitioned formula data to form a formula distribution spectrum;

[0031] S305. Perform data regression analysis on each sampling point using the recipe distribution spectrum to obtain recipe data;

[0032] S306. Match and analyze the formula data with the performance index range, and output optimized formula materials.

[0033] The forming analysis phase of step S4 includes the following steps:

[0034] S401. Collecting the temperature field and pressure field of the optimized formula material at multiple points, and obtaining the molding process data by comparative analysis of the temperature field and pressure field gradients;

[0035] S402. Calculate the time evolution characteristics of material stress distribution, deformation and density distribution according to the molding process data to generate molding process parameters;

[0036] S403. The forming process parameters are numerically discretely calculated according to 1024 grid nodes to obtain partition forming parameters;

[0037] S404. Perform multi-field coupling calculation on the temperature field, pressure field and deformation field in the partition molding parameters to form a molding state curve;

[0038] S405. Perform trend analysis on each monitoring point using the molding state curve to obtain molding process parameters;

[0039] S406: Perform deviation analysis on the molding process parameters and process thresholds, and output molding materials.

[0040] The curing analysis stage of step S5 includes the following steps:

[0041] S501. Performing time series acquisition of the temperature field and infrared radiation intensity monitoring at multiple measuring points on the molding material, and obtaining a curing state sequence through spatiotemporal evolution analysis of the temperature field distribution and the radiation intensity gradient;

[0042] S502. Calculate the dynamic change characteristics of the crosslinking degree distribution, the curing depth and the stress distribution according to the curing state sequence, and generate curing process parameters;

[0043] S503. The curing process parameters are classified and processed according to 2048 sampling points to obtain partition curing parameters;

[0044] S504. Perform multi-field collaborative calculation on the temperature field evolution, stress distribution and curing degree in the partition curing parameters to form a curing dynamic spectrum;

[0045] S505. Using the curing dynamic spectrum, data reconstruction and analysis are performed on the curing degree of each sensor node to obtain curing process parameters;

[0046] S506. Perform process trajectory planning and curve fitting analysis on the curing process parameters, and output a cured product.

[0047] The quality evaluation stage in step S6 includes the following steps:

[0048] S601. Perform multi-dimensional physical property index detection and internal structure scanning on the cured product, and obtain a quality feature matrix through correlation analysis between physical property parameter distribution and internal defect characteristics;

[0049] S602. Calculate the spatial distribution law of intensity distribution, thermal conductivity and density uniformity according to the quality characteristic matrix, and generate evaluation parameters;

[0050] S603. Perform multi-scale decomposition processing on the evaluation parameters according to 4096 feature points to obtain partition evaluation parameters;

[0051] S604. Perform data fusion calculation on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters to form a quality evaluation surface;

[0052] S605. Perform sensitivity analysis on the abnormality of each test item using the quality evaluation surface to obtain evaluation parameters;

[0053] S606. Perform multi-objective optimization and parameter mapping analysis on the evaluation parameters and process control variables, and output a production control plan.

[0054] In step S604, data fusion calculation is performed on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters to form a quality evaluation surface, including the following steps:

[0055] S6041. Normalizing and calculating the standard deviation of the physical performance data in the partition evaluation parameters, and obtaining a physical property fusion matrix through weight allocation and data dimensionality reduction analysis;

[0056] S6042. Perform cross validation and correlation analysis based on the physical property fusion matrix and mechanical performance data to generate a mechanical feature sequence;

[0057] S6043. Reconstruct the mechanical feature sequence and the structural feature data according to 8192 sampling points to obtain feature fusion parameters;

[0058] S6044. Perform principal component extraction and singular value decomposition calculation on the feature fusion parameters to form a feature mapping matrix;

[0059] S6045. Perform multidimensional data space transformation and coordinate mapping processing using the feature mapping matrix to obtain surface construction parameters;

[0060] S6046. Perform isosurface fitting and boundary continuity constraint analysis using the surface construction parameters, and output a quality evaluation surface.

[0061] In a second aspect, the present application provides a graphite composite insulation board production process control system, the graphite composite insulation board production process control system comprising:

[0062] An analysis module is used to analyze and process the raw materials to obtain raw material characteristic data, and optimize the ratio parameters based on the raw material characteristic data to obtain pre-processed raw materials;

[0063] A monitoring module, used to monitor and process the state parameters corresponding to the pre-treated raw material to obtain reaction state data, and use the reaction state data to control the process parameters to form a modified raw material;

[0064] An optimization module, used to perform a formulation optimization process according to the material parameters of the modified raw material, obtain formulation data, complete a material performance prediction process according to the formulation data, and generate an optimized formulation material;

[0065] A control module, used for performing molding analysis processing through the physical parameters of the optimized formula material to obtain molding process parameters, and implementing process control processing based on the molding process parameters to produce molding materials;

[0066] A processing module, used for performing curing analysis processing using the state parameters of the molding material to obtain curing process parameters, and performing curing process control processing according to the curing process parameters to form a cured product;

[0067] A generation module is used to perform quality evaluation processing based on the inspection data of the cured product, obtain evaluation parameters, execute control strategy generation processing according to the evaluation parameters, and generate a production control plan.

[0068] A third aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned graphite composite insulation board production process control method.

[0069] The beneficial effects of the present invention are:

[0070] 1. Through the analysis and processing of raw materials, the characteristics of raw materials are accurately obtained and the proportion parameters are optimized, which ensures the quality stability of pre-treated raw materials and provides a reliable raw material basis for subsequent processing;

[0071] 2. According to the state parameters corresponding to the pre-treated raw materials, monitoring and processing are carried out, and a real-time monitoring mechanism for reaction state data is established. Through the dynamic regulation of process parameters, the controllability of the modification process is ensured, and the performance consistency of the modified raw materials is improved;

[0072] 3. Formula optimization is carried out according to the material parameters of the modified raw materials, which realizes the scientific management of formula data. Through material performance prediction processing, the performance predictability of the optimized formula materials is guaranteed;

[0073] 4. By optimizing the physical parameters of the formula materials and performing molding analysis and processing, a precise control system for molding process parameters was established. Based on the process control of molding process parameters, the quality level of molding materials was improved;

[0074] 5. The state parameters of the molding material are used for curing analysis and processing, which realizes the intelligent control of the curing process. The performance stability of the cured product is ensured through the precise regulation of the curing process;

[0075] 6. Based on the test data of cured products, quality evaluation is carried out and a complete quality evaluation system is established. Through the optimization and generation of control strategies, closed-loop control of the production process is achieved, which significantly improves the consistency of product quality and production efficiency;

[0076] 7. The present invention realizes intelligent control of the entire process from raw materials to finished products by establishing a multi-dimensional data collection, processing and control system, effectively solving the problems existing in traditional production methods such as large quality fluctuations, insufficient optimization of process parameters, and low process control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the drawings required for use in the description of the embodiment will be briefly introduced below. Obviously, the drawings in the following description are embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0078] Figure 1 This is a flow chart of a method for controlling the production process of a graphite composite thermal insulation board in an embodiment of the present application;

[0079] Figure 2 This is a schematic diagram of the control system for the production process of graphite composite insulation board in the embodiment of this application. DETAILED DESCRIPTION

[0080] The embodiments of the present application provide a graphite composite insulation board production process control method, control system and storage medium. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0081] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1In an embodiment of the present application, a method for controlling the production process of a graphite composite insulation board comprises the following steps:

[0082] It is understandable that the execution subject of the present application may be a graphite composite insulation board production process control system, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0083] Step S1, raw material analysis stage: analyze and process the raw materials to obtain raw material characteristic data, optimize the ratio parameters based on the raw material characteristic data, and obtain pre-processed raw materials; analyze and process the raw materials, and collect the morphological image data of the raw materials through a multi-spectral camera. These image data contain information such as the morphological characteristics, particle size distribution, and surface texture of the raw materials. The acquisition of raw material characteristic data is achieved by digital conversion, feature extraction, and analysis of the collected images. The original image is grayed to obtain a grayscale image, and the grayscale image is binarized to obtain a binary image. Then, the binary image is subjected to edge detection, morphological processing, and other operations to extract the morphological characteristic parameters of the raw materials. These parameters include the average size, shape factor, and degree of dispersion of the particles.

[0084] Step S2, monitoring and processing stage: monitor and process according to the state parameters corresponding to the pre-treated raw materials to obtain reaction state data, use the reaction state data to control the process parameters, and form modified raw materials; in this stage, the pre-treated raw materials are subjected to multi-time temperature field acquisition. The temperature field acquisition is realized by an infrared thermal imager, the acquisition interval is 0.5 seconds, and the temperature value is recorded at each sampling point. After the temperature field data is processed by digital filtering and denoising, the temperature gradient distribution is calculated. The temperature gradient distribution data is used to characterize the uniformity and stability of the reaction. The reaction state data is obtained by performing time series analysis and feature extraction on the temperature field data, including parameters such as reaction rate, conversion rate and degree of dispersion.

[0085] Step S3, formula optimization stage: perform formula optimization processing according to the material parameters of the modified raw materials to obtain formula data, complete material performance prediction processing according to the formula data, and generate optimized formula materials; at this stage, the material parameters of the modified raw materials are basic data obtained through physical testing, including thermal conductivity, density, strength, etc. After these data are normalized, comprehensive evaluation indicators are obtained through data fusion algorithms. The process of generating formula data is to correlate material parameters with process parameters and establish a mapping relationship between parameters. Material performance prediction is based on existing data samples, and a performance prediction model is established through regression analysis methods.

[0086] Step S4, molding analysis stage: perform molding analysis processing by optimizing the physical parameters of the formula materials, obtain molding process parameters, implement process control processing based on the molding process parameters, and produce molding materials; in this stage, it involves testing the physical parameters of the optimized formula materials, including parameters such as fluidity and thixotropy. The molding process parameters are obtained by analyzing the physical parameters, including molding pressure, temperature, time and other process parameters. The process control processing is the process of adjusting the process conditions in real time according to the molding process parameters.

[0087] Step S5, curing analysis stage: the state parameters of the molding material are used for curing analysis and processing to obtain curing process parameters, and curing process control processing is implemented according to the curing process parameters to form a cured product; In this stage, the state parameters of the molding material need to be monitored, including parameters such as temperature and hardness. The curing process parameters are obtained by analyzing and processing the state parameters, including parameters such as curing degree and cross-linking degree. Process control processing is the process of adjusting the curing process conditions according to the curing process parameters.

[0088] Step S6, quality evaluation stage: perform quality evaluation based on the test data of the cured product, obtain evaluation parameters, perform control strategy generation based on the evaluation parameters, and generate a production control plan. In this stage, a comprehensive performance test is performed on the cured product to obtain test data. The evaluation parameters are comprehensive performance indicators obtained by analyzing and processing the test data. The control strategy is generated based on the evaluation parameters, and the optimal process parameter combination is determined by the optimization algorithm.

[0089] The data processing in the raw material analysis stage in step S1 focuses on feature extraction and parameter optimization. In the monitoring and processing stage in step S2, data processing mainly includes signal filtering, feature extraction and state recognition. The data processing in the formulation optimization stage in step S3 focuses on multi-parameter fusion and performance prediction. The data processing in the molding analysis stage in step S4 is mainly process parameter optimization and process control. The data processing in the curing analysis stage in step S5 includes state monitoring and process optimization. The data processing in the quality evaluation stage in step S6 focuses on comprehensive performance evaluation and control strategy optimization.

[0090] For example, in the raw material analysis stage of step S1, a multispectral image of graphite raw material is first collected, and the image resolution is 4096×3072 pixels. After the image is grayed, it is binarized by the local adaptive threshold method to obtain the contour image of the particles. The particle features are extracted by morphological processing, and the average particle size and distribution uniformity are calculated. In the monitoring and processing stage of step S2, the temperature field of the pretreated raw material is collected, with a sampling frequency of 2Hz and a collection time of 30 minutes. The temperature data is processed by a Butterworth low-pass filter with a cutoff frequency of 0.5Hz to obtain the filtered temperature field data. In the formulation optimization stage of step S3, the thermal conductivity of the material is tested, and the transient plane heat source method is used, and the test temperature range is 25-80℃. The test data is subjected to linear regression analysis to establish a relationship model between thermal conductivity and temperature. In the molding analysis stage of step S4, the rheological properties of the material are measured, and a rotational rheometer is used to determine the molding process parameters through the viscosity-shear rate curve. In the curing analysis stage of step S5, the curing characteristics of the material are determined by differential scanning calorimetry, with a heating rate of 10°C / min and a temperature range of 30-200°C. The degree of curing and reaction rate are calculated based on the heat flow curve. In the quality evaluation stage of step S6, the thermal conductivity, compressive strength and density of the product are tested, and comprehensive performance indicators are obtained through data fusion. According to the deviation between the performance indicators and the target values, the process parameters are optimized and adjusted.

[0091] In the embodiments of the present application, by analyzing and processing the raw materials, the characteristics of the raw materials are accurately acquired and the proportioning parameters are optimized, the quality stability of the pre-treated raw materials is guaranteed, and a reliable raw material basis is provided for subsequent processing; monitoring and processing are performed according to the state parameters corresponding to the pre-treated raw materials, and a real-time monitoring mechanism for the reaction state data is established. Through the dynamic regulation of the process parameters, the controllability of the modification process is ensured, and the performance consistency of the modified raw materials is improved; the formula optimization processing is performed according to the material parameters of the modified raw materials, and the scientific management of the formula data is realized. Through the material performance prediction processing, the performance predictability of the optimized formula material is guaranteed; the molding analysis processing is performed by optimizing the physical parameters of the formula materials, and an accurate control system of the molding process parameters is established. The process regulation based on the molding process parameters improves the quality level of the molding materials; the state parameters of the molding materials are used for curing analysis processing, and the intelligent control of the curing process is realized. Through the precise regulation of the curing process, the performance stability of the cured product is guaranteed; quality evaluation processing is performed based on the detection data of the cured product, and a complete quality evaluation system is established. Through the optimization generation of the control strategy, the closed-loop control of the production process is realized, and the consistency of product quality and production efficiency are significantly improved. The present invention realizes intelligent control of the entire process from raw materials to finished products by establishing a multi-dimensional data collection, processing and control system, effectively solving the problems existing in traditional production methods such as large quality fluctuations, insufficient optimization of process parameters, and low process control accuracy.

[0092] In a specific embodiment, the process of executing step S1 raw material analysis stage may specifically include the following steps:

[0093] S101. Perform multi-time spectrum acquisition on the raw material, and obtain the raw material spectrum characteristic data by comparing the spectrum intensity with the standard spectrum; specifically, perform multi-time spectrum acquisition on the raw material through the spectrum acquisition equipment, the acquisition process lasts for 120 minutes, and the spectrum data is collected every 10 minutes. The wavelength range of the spectrum acquisition covers 200-2500nm, and the sampling interval is 1nm. The collected spectrum data is compared with the pre-established standard spectrum database, which contains the standard spectrum characteristics of raw materials of different quality grades. The spectral characteristic data of the raw material is obtained by calculating the correlation coefficient and Euclidean distance between the collected spectrum and the standard spectrum.

[0094] S102. Calculate the time period distribution of physical and chemical characteristics, particle size distribution and morphological characteristics based on the raw material spectral characteristic data to generate initial characteristic standard data; physical characteristics include parameters such as density, particle size, specific surface area, chemical characteristics include indicators such as carbon content and impurity content, and morphological characteristics include parameters such as particle shape factor and surface roughness. Arrange these characteristic parameters in time series to construct a time period distribution matrix. The time period distribution matrix records the change pattern of the raw material characteristic parameters over time, and generates initial characteristic standard data by standardizing the matrix.

[0095] S103. The initial characteristic standard data is partitioned and calculated according to 64 data channels to obtain partition characteristic data; the data channel division is based on the spectral band and feature type, where 16 channels are allocated to the visible light region, 24 channels are allocated to the near infrared region, and 24 channels are allocated to the mid-infrared region. The data in each channel is independently normalized and denoised, and high-frequency noise is eliminated by wavelet transform to retain effective signal characteristics. After partition calculation, the partition characteristic data of each channel is obtained.

[0096] S104. The physical features, chemical features and morphological features in the partition feature data are cumulatively calculated to form a feature cumulative spectrum; that is, the physical features, chemical features and morphological features are accumulated in the time dimension. The cumulative calculation adopts a weighted average method, and the weight coefficient is dynamically adjusted according to the importance of the feature. The result of the cumulative calculation is presented in the form of a three-dimensional spectrum, that is, a feature cumulative spectrum, the horizontal axis represents the data channel, the vertical axis represents the feature type, and the color depth represents the cumulative value. The distribution law of the raw material characteristics can be intuitively reflected through the spectrum.

[0097] S105. Use the characteristic cumulative spectrum to perform distribution calculation on each data channel of the raw material to obtain raw material characteristic data; the distribution calculation includes statistical parameters such as normality test, kurtosis calculation, skewness analysis, etc. By calculating the data distribution characteristics of each channel and establishing a characteristic space mapping relationship, a raw material characteristic data set that can fully characterize the characteristics of the raw material is finally obtained.

[0098] S106. Perform correlation analysis on the raw material characteristic data and the ratio optimization parameters, and output the pre-treated raw materials. The ratio optimization parameters include the mass ratio of each component, the mixing order, the mixing time and other process parameters. Through correlation analysis and principal component analysis, the mapping relationship between the characteristic parameters and the ratio parameters is determined, and a ratio optimization model is established. According to the calculation results of the optimization model, the optimal ratio scheme is determined, and the process parameters of the pre-treated raw materials are output.

[0099] In a specific embodiment, the process of executing step S2 monitoring processing stage may specifically include the following steps:

[0100] S201. Collect the temperature field of the pretreated raw materials in multiple time periods, and obtain the temperature change data by comparing the temperature field intensity with the standard temperature field; specifically, the temperature field is collected by real-time scanning with an infrared thermal imager, and the scanning range covers the entire reaction area with a resolution of 0.1°C. During the collection process, the complete temperature field data is recorded every 5 minutes for 4 hours. The collected temperature field data is compared with the pre-set standard temperature field, which is an optimal temperature distribution pattern established based on a large amount of historical production data statistics. By calculating the difference between the measured temperature field and the standard temperature field, the time series data reflecting the temperature change trend is obtained.

[0101] S202. Calculate the time distribution characteristics of reaction rate, conversion rate and dispersion degree according to the temperature change data, and generate reaction state data; temperature change data is the basis for calculating reaction kinetic parameters. The reaction rate is calculated by the temperature change rate, which reflects the temperature change per unit time. The conversion rate characterizes the reaction process and is calculated by the ratio of the real-time temperature to the theoretical maximum temperature. The dispersion degree is represented by the uniformity coefficient of the temperature field, and the larger the coefficient, the more uniform the temperature distribution. These parameters together constitute the reaction state data, which are arranged in chronological order to construct the time distribution characteristic matrix of the reaction process, which comprehensively records the dynamic change characteristics of the reaction process.

[0102] S203. The reaction state data is divided into regions and calculated according to 256 data sampling points to obtain partitioned reaction state data; the sampling points of the reaction state data are set by an equidistant distribution method, and 256 virtual sampling points are arranged in the reactor. The spatial distribution of these sampling points takes into account the geometric characteristics of the reactor and the variation law of the temperature field. Each sampling point independently records the three key parameters of reaction rate, conversion rate and dispersion degree. By dividing the data of these sampling points into regions, a mathematical description of the local reaction state is established to obtain partitioned reaction state data.

[0103] S204. Perform weighted cumulative calculation on the reaction rate, conversion rate and dispersion degree in the partitioned reaction state data to form a state cumulative distribution; when performing weighted cumulative calculation on the partitioned reaction state data, an adaptive weight allocation method is adopted. The weight coefficient is dynamically adjusted according to the importance of the parameter and the reliability of the data. The reaction rate weight coefficient is 0.4, the conversion rate weight coefficient is 0.4, and the dispersion degree weight coefficient is 0.2. Through weighted cumulative operation, a state cumulative distribution is formed and presented in the form of a cumulative distribution curve, which clearly shows the overall reaction state and evolution law of the reaction process.

[0104] S205. The state parameters of each sampling point are subjected to distribution fitting calculation using the state cumulative distribution to obtain reaction state data; the state cumulative distribution is the data basis for distribution fitting calculation. Mathematical statistical analysis is performed on the state parameters of each sampling point, including calculation of statistical quantities such as mean, variance, skewness and kurtosis. These statistical parameters are fitted by the least squares method to establish a distribution model of the state parameters. The fitting calculation results form a complete reaction state data set, which accurately describes the various characteristic parameters of the reaction process.

[0105] S206. Compare and analyze the reaction state data with the process parameter threshold range, and output the modified raw material. Compare the reaction state data with the preset process parameter threshold range. The process parameter threshold is a control range determined according to the product quality requirements, including temperature threshold, pressure threshold, time threshold, etc. Through multi-dimensional parameter comparison, it is determined whether the reaction state meets the process requirements. When all parameters are within the threshold range, it can be confirmed that the modification process is completed and the modified raw material is output.

[0106] In a specific embodiment, the process of executing step S3 formulation optimization stage may specifically include the following steps:

[0107] S301. Collect the physical parameters of the modified raw materials in multiple time periods, and obtain the material parameter sequence through comparative analysis of thermal conductivity, density and strength; specifically, the collection of physical parameters of modified raw materials is a systematic data acquisition process. The physical parameters include three core indicators: thermal conductivity, density and strength. Thermal conductivity is measured by the transient plane heat source method. The measurement temperature range is from room temperature to 200°C, and a data point is recorded every 20°C. The density is determined by the Archimedes drainage method, and each sample is measured 5 times to take the average value. The strength test uses a universal testing machine to measure the compressive strength, and the loading rate is 2mm / min. By establishing a standard test database, the measured data is compared and analyzed with the standard data to obtain the material parameter sequence.

[0108] The processing of material parameter sequences involves several mathematical models. For numerical decomposition calculations, the following mathematical models are used:

[0109] ;

[0110] in, represents the decomposed numerical distribution function, r is the spatial coordinate, t is the time variable, and is the weight coefficient, is the spatial basis function, is the time basis function, is the coupling term, n and m are the decomposition orders in space and time dimensions respectively.

[0111] S302. Calculate the time variation characteristics of material strength, thermal conductivity and density distribution according to the material parameter sequence to generate formula data; calculate the variation characteristics of material performance indicators based on the material parameter sequence. The time correlation of strength variation is obtained by linear regression analysis. The variation of thermal conductivity is analyzed by Fourier heat conduction model. The density distribution is fitted by Gaussian mixture model. These time variation characteristics are combined to form complete formula data.

[0112] S303. The formula data is numerically decomposed and calculated according to 512 sampling points to obtain partitioned formula data; the formula data is processed using 512 sampling points for refined analysis. The arrangement of the sampling points takes into account the anisotropy of the material and is evenly distributed in three-dimensional space. Each sampling point simultaneously records three types of indicators: physical properties, mechanical properties, and thermal conductivity.

[0113] S304. Perform weighted statistical calculations on the physical properties, mechanical properties and thermal conductivity in the partitioned formula data to form a formula distribution spectrum; physical properties include parameters such as density and porosity, mechanical properties include parameters such as compressive strength and elastic modulus, and thermal conductivity includes parameters such as thermal conductivity and thermal diffusion coefficient. The weighted statistical calculation of the partitioned formula data uses the analytic hierarchy process to determine the weights. The weight ratio of physical properties, mechanical properties and thermal conductivity is determined by the expert scoring method. The parameter weights within each performance indicator are calculated by the principal component analysis method. The weighted statistical process forms a three-dimensional formula distribution spectrum, and the three coordinate axes of the spectrum correspond to three types of performance indicators, and the color depth represents the comprehensive performance score.

[0114] S305. Perform data regression analysis on each sampling point using the recipe distribution spectrum to obtain recipe data; the data regression analysis of the recipe distribution spectrum uses a multivariate nonlinear regression method. The regression model includes three forms: polynomial regression, exponential regression, and logarithmic regression. The optimal model is selected through residual analysis. The results of the regression analysis form recipe data, which reflect the quantitative relationship between material properties and recipe parameters.

[0115] S306. Match and analyze the formula data with the performance index range, and output the optimized formula material.

[0116] The matching analysis of performance indicators adopts fuzzy comprehensive evaluation method. An evaluation index system is established, including three primary indicators of thermal conductivity, mechanical properties, and economic performance, and multiple secondary indicators. The comprehensive score is obtained through fuzzy evaluation matrix calculation, and the score result is used as the basis for selecting the optimized formula.

[0117] For example, when collecting physical parameters of modified raw materials, the samples are first cut into standard sizes for pretreatment. In the thermal conductivity test, the test current is set to 10mA, the heating time is 10s, and the temperature response curve is collected. In the density test, the standard density bottle method is used to control the test temperature at 25±0.1℃. The strength test uses a standard specimen, and the pressurization speed is controlled at 0.5MPa / s. The collected data is standardized to eliminate the dimension effect. The correlation between the parameters is determined through correlation analysis. Strength and density show a positive correlation, and thermal conductivity and density also have a significant correlation. Based on these correlations, a performance prediction model is established. The model is verified by cross-validation method, and the predicted results are in good agreement with the measured data. The formula data of each sampling point is processed by data fusion to form a complete data matrix. The data matrix reflects the distribution law of material properties in space and time.

[0118] In a specific embodiment, the process of executing step S4 of the forming analysis phase may specifically include the following steps:

[0119] S401. Multi-point temperature field and pressure field acquisition is performed on the optimized formula material, and the molding process data is obtained by comparative analysis of the temperature field and pressure field gradients; specifically, the distributed sensing technology is used for multi-point monitoring of the optimized formula material, and the temperature sensor array and the pressure sensor array are arranged on the surface of the molding mold. The temperature sensor adopts a K-type thermocouple with a measurement range of 0-300℃ and an accuracy of ±0.1℃; the pressure sensor adopts a piezoresistive sensor with a range of 0-20MPa and an accuracy of ±0.1%FS. The temperature field and pressure field data are synchronously collected through a high-speed data acquisition system with a sampling frequency of 100Hz. The comparative analysis of the temperature field and pressure field gradients adopts a spatial difference algorithm to calculate the gradient values ​​between each monitoring point, obtain the molding process data, and form a gradient distribution matrix. In the comparative analysis process, the actual gradient value is compared with the predicted value of the theoretical model to obtain the evolution law of the temperature field and pressure field in the molding process.

[0120] S402. Calculate the time evolution characteristics of material stress distribution, deformation and density distribution according to the molding process data, and generate molding process parameters; establish a material mechanical properties analysis model based on the acquired molding process data. The stress distribution is calculated by elastic-plastic mechanics theory, considering the influence of temperature field on the mechanical properties of the material. The deformation is calculated based on the displacement field data, and the deformation equation is solved by numerical integration method. The density distribution is calculated by the principle of mass conservation combined with the deformation field data. The time evolution characteristics of these physical quantities are extracted by time series analysis methods, including trend analysis, periodicity analysis and mutation feature analysis. By constructing a time series model, the change trend of each physical quantity is predicted and the molding process parameters are generated.

[0121] The S403 molding process parameters are numerically discretely calculated according to 1024 grid nodes to obtain the partition molding parameters; the spatial discretization of the molding process parameters uses 1024 grid nodes, and the grid division uses adaptive grid technology, and the grid is encrypted in the stress concentration area and the area with large temperature gradient. Each grid node records parameters such as temperature, pressure, stress, deformation and density. The finite element method is used to numerically solve each physical field to obtain the partition molding parameters. The calculation of the partition parameters uses parallel computing technology to improve the calculation efficiency.

[0122] S404. Perform multi-field coupling calculation on the temperature field, pressure field and deformation field in the partition molding parameters to form a molding state curve;

[0123] Multi-field coupling calculation involves the interaction of temperature field, pressure field and deformation field. Its mathematical model is expressed as follows:

[0124] ;

[0125] in, is the coupled field function, (x, y, z) is the spatial coordinate, t is the time variable, , , They are the weight coefficients of temperature field, pressure field and deformation field respectively. , , are the basis functions of temperature field, pressure field and deformation field respectively, is the coupling coefficient, is the field coupling term, n, m, l are the number of expanded terms of temperature field, pressure field and deformation field respectively.

[0126] The coupled calculation results form a forming state curve, which reflects the dynamic changes of the material state during the forming process.

[0127] S405. Perform trend analysis on each monitoring point using the molding state curve to obtain molding process parameters;

[0128] The trend analysis of each monitoring point adopts time series analysis methods, including moving average, exponential smoothing and autoregressive analysis, etc. Through trend analysis, the key time points and abnormal conditions in the molding process are identified, and the molding process control parameters are generated.

[0129] S406. Perform deviation analysis on molding process parameters and process thresholds, and output molding materials.

[0130] The deviation analysis between molding process parameters and process thresholds adopts fuzzy control theory. A fuzzy rule base is established to map the deviation values ​​of temperature, pressure, stress and other parameters to the adjustment values ​​of the control quantity. The adjustment plan of process parameters is obtained through fuzzy reasoning calculation to ensure that the molding process is carried out within the optimal process range.

[0131] For example, in the monitoring of the molding process, the sensor layout of the mold is first designed. The temperature sensors are evenly arranged on the mold surface with a spacing of 50mm to form a monitoring network. The pressure sensors are arranged at key stress positions to monitor the pressure distribution. During the data acquisition process, the temperature field and pressure field data are recorded synchronously. The collected data is subjected to noise reduction to eliminate the influence of random disturbances. The spatial distribution characteristics of the temperature field and pressure field are calculated through gradient analysis. The temperature gradient reflects the direction and intensity of heat transfer, and the pressure gradient reflects the trend of material flow. Based on this gradient information, a mathematical model of the molding process is established. The model takes into account factors such as thermal deformation, compression deformation and flow deformation of the material.

[0132] In the grid discretization calculation, the tetrahedral grid is used to divide the molding space. The grid size is automatically adjusted according to the local parameter change rate to ensure the calculation accuracy. The heat conduction equation, stress balance equation and continuity equation are solved at each grid node to obtain the complete field quantity distribution. In the multi-field coupling calculation, the influence of temperature on material properties, the influence of pressure on deformation, and the influence of deformation on the temperature field and pressure field distribution are considered. The generation of the molding state curve is based on the comprehensive analysis of multiple physical quantities. The curve reflects the state evolution of the material during the molding process, including information such as temperature change, density change, and stress change. By analyzing the characteristic points and change trends of the curve, the key stages of the molding process can be identified. This information is used to guide the adjustment of process parameters to ensure molding quality.

[0133] In a specific embodiment, the process of executing step S5 of the curing analysis phase may specifically include the following steps:

[0134] S501. Perform time-series acquisition of the temperature field at multiple measuring points and monitor the infrared radiation intensity of the molding material, and obtain the curing state sequence through the spatiotemporal evolution analysis of the temperature field distribution and the radiation intensity gradient; specifically, the time-series acquisition of the temperature field at multiple measuring points adopts distributed fiber optic sensing technology, and a fiber optic sensor network is arranged in the curing furnace to achieve high-precision real-time monitoring of the temperature field. The infrared radiation intensity monitoring adopts a planar array infrared detector to collect the infrared radiation characteristics of the material surface during the curing process. The temperature field distribution data is obtained by a fiber grating demodulator with a sampling frequency of 1kHz and a spatial resolution of 1mm. The infrared radiation intensity data is collected by an infrared imaging system with a wavelength range of 3-5μm and a temperature resolution of 0.1℃. Through the spatiotemporal evolution analysis algorithm, the temperature field data and the infrared radiation data are spatiotemporally aligned to establish a curing state sequence. The curing state sequence is the data basis for calculating the crosslinking degree distribution, curing depth and stress distribution.

[0135] S502. The dynamic change characteristics of the cross-linking degree distribution, curing depth and stress distribution are calculated according to the curing state sequence to generate the curing process parameters; the cross-linking degree distribution is calculated by thermal analysis method based on temperature field data and the thermal curing kinetic model of the material. The curing depth is calculated based on the attenuation characteristics of the infrared radiation intensity, reflecting the degree of advancement of the curing reaction from the surface to the inside. The stress distribution is obtained by thermal-mechanical coupling analysis, taking into account the thermal stress and chemical shrinkage stress caused by the temperature gradient. The dynamic change characteristics of these physical quantities are extracted by time series analysis method to form a complete set of curing process parameters.

[0136] S503. The curing process parameters are classified and processed according to 2048 sampling points to obtain partitioned curing parameters; the spatial discretization of the curing process parameters uses 2048 sampling points, and the distribution of the sampling points takes into account the changing characteristics of the temperature field and the stress field. In areas with large temperature gradients and stress concentration areas, the density of sampling points increases accordingly. Each sampling point records parameters such as temperature, stress, displacement and degree of curing. These parameters are classified by cluster analysis method to obtain partitioned curing parameters. The classification process takes into account the correlation and spatial distribution characteristics between parameters.

[0137] S504. Perform multi-field collaborative calculation on the temperature field evolution, stress distribution and curing degree in the partition curing parameters to form a curing dynamic spectrum;

[0138] Multi-field collaborative calculation involves the coupling effect of temperature field, stress field and curing degree field. Its mathematical model is expressed as follows:

[0139] ;

[0140] in, is the synergistic field function, (x, y, z) is the spatial coordinate, t is the time variable, , , are the weight coefficients of temperature field, stress field and curing degree field, respectively. , , are the basis functions of temperature field, stress field and curing degree field respectively, is the coupling coefficient, is the field coupling term, p, q, r are the number of expansion terms of temperature field, stress field and curing degree field respectively.

[0141] The collaborative calculation results form a solidification dynamic spectrum, which reflects the evolution laws of various physical fields during the solidification process.

[0142] S505. Use the curing dynamic spectrum to reconstruct the data of the curing degree of each sensor node to obtain the curing process parameters; perform data reconstruction analysis on the curing degree of each sensor node, use the Kriging interpolation method to perform spatial reconstruction, and obtain a continuous curing degree distribution field. The reconstruction process takes into account the spatial correlation of the measuring points and the reliability of the data to generate complete curing process parameters.

[0143] S506. Perform process trajectory planning and curve fitting analysis on the curing process parameters and output the cured product. The process trajectory planning is based on the curing process parameters and uses a dynamic programming method to determine the optimal curing process path. The process trajectory includes three dimensions: temperature curve, pressure curve, and time curve. The process curve is smoothed by a polynomial fitting method to ensure the continuity and controllability of the process parameters. The final output cured product meets the various performance indicators required by the design.

[0144] For example, a batch of graphite composite insulation boards with specifications of 1000mm×600mm×50mm are controlled during the curing process. First, a sensor network is arranged. The temperature sensor is arranged in a layer every 10mm in the thickness direction, and 81 measuring points are evenly distributed on each layer. The infrared detector is placed on the top of the curing furnace, and the field of view ensures that the entire product surface is covered. During the data collection process, the temperature values ​​of each layer of measuring points and the surface infrared radiation intensity are recorded in real time. The temperature data is digitally filtered to eliminate noise, and the infrared data is corrected by atmospheric transmittance. The corresponding relationship between the temperature field and the infrared field is established through the spatiotemporal registration algorithm. Based on these data, the degree of progress of the curing reaction is calculated.

[0145] The degree of cure is calculated using the DSC method, and the reaction conversion rate is determined by thermal analysis. The stress distribution is calculated using the finite element method, taking into account the coupling effect of thermal stress and chemical shrinkage stress. The partition parameters are calculated using the K-means clustering algorithm, and regions with similar characteristics are grouped together. The parameters of each partition are calculated and controlled independently, achieving precise control of the partition. A complete curing dynamic spectrum is obtained through multi-field collaborative calculation. The spectrum shows the spatiotemporal evolution of the temperature field, stress field, and degree of cure field. Based on the spectrum analysis, key process parameters are determined, including heating rate, holding time, and cooling rate. The planning of the process trajectory takes into account the curing kinetics of the material and the process constraints to ensure the controllability and repeatability of the curing process.

[0146] In a specific embodiment, the process of executing step S6 quality evaluation stage may specifically include the following steps:

[0147] S601. Perform multi-dimensional physical property index detection and internal structure scanning on the cured product, and obtain the quality feature matrix through correlation analysis between physical property parameter distribution and internal defect characteristics;

[0148] Specifically, the multi-dimensional physical property index detection adopts a method combining multiple detection methods, including thermal conductivity test, density test, strength test, etc. The internal structure scanning adopts industrial CT technology to obtain the three-dimensional structural information inside the product. The thermal conductivity test adopts the transient hot wire method, and the measuring points are evenly arranged on the surface of the product. Each measuring point records the thermal diffusion coefficient and thermal conductivity. The density test adopts the Archimedes drainage method, and samples are tested at different positions to obtain density distribution data. The strength test adopts the compression test method to test the compressive strength and elastic modulus. CT scanning uses high-resolution industrial CT, and the layer spacing is set to 0.1mm to obtain a clear internal structure image. Through the correlation analysis of the distribution of physical property parameters and internal defect characteristics, the parameter-defect mapping relationship is established. The correlation analysis adopts multivariate statistical methods to calculate the correlation coefficient and partial correlation coefficient between each parameter. Image processing and feature extraction are performed on the CT image to identify internal defects such as bubbles and cracks. The physical property parameters and defect characteristics are combined into a quality feature matrix, and each element in the matrix represents the comprehensive quality characteristics of a measuring point.

[0149] S602. Calculate the spatial distribution of strength distribution, thermal conductivity and density uniformity according to the quality characteristic matrix, and generate evaluation parameters; the quality characteristic matrix is ​​the data basis for calculating strength distribution, thermal conductivity and density uniformity. The strength distribution is calculated by mechanical analysis method, considering the influence of internal defects on strength. The spatial distribution of thermal conductivity is analyzed by heat conduction theory, and a heat conduction model is established. Density uniformity is calculated by statistical analysis method, including variance analysis and coefficient of variation calculation. These distribution laws form a complete spatial distribution feature to generate evaluation parameters.

[0150] S603. The evaluation parameters are decomposed at multiple scales according to 4096 feature points to obtain partition evaluation parameters. The spatial discretization of the evaluation parameters uses 4096 feature points, and the distribution of feature points takes into account the geometric characteristics of the product and the law of performance change. In areas of performance mutation and defect-intensive areas, the density of feature points increases accordingly. Each feature point records multidimensional parameters such as physical properties, mechanical properties and structural characteristics. Multi-scale decomposition is performed through wavelet transform to obtain feature performance at different scales and form partition evaluation parameters.

[0151] S604. Perform data fusion calculation on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters to form a quality evaluation surface; the data fusion of physical properties, mechanical properties and structural characteristics is the core link of quality evaluation. The fusion calculation adopts a multi-level fusion model, including data layer fusion, feature layer fusion and decision layer fusion. The data layer fusion mainly deals with the normalization and standardization of the original test data. The feature layer fusion performs feature extraction and selection to screen out the most representative feature parameters. The decision layer fusion adopts a fuzzy comprehensive evaluation method. The final quality evaluation surface is obtained by combining various indicators.

[0152] S605. Use the quality evaluation surface to perform sensitivity analysis on the abnormality of each test item to obtain evaluation parameters; identify the abnormality of each test item by analyzing the quality evaluation surface. The determination of the abnormality is based on statistical process control theory, and control limits and alarm limits are established. Focus on analyzing the parameters that exceed the control limits to determine the cause and impact of the abnormality. Generate a complete set of evaluation parameters based on the abnormality analysis results.

[0153] S606. Perform multi-objective optimization and parameter mapping analysis on the evaluation parameters and process control variables, and output the production control plan. The production control plan is generated using a multi-objective optimization method. Multi-objective optimization should consider multiple objectives such as product quality, production efficiency and cost at the same time. Parameter mapping analysis establishes the corresponding relationship between process parameters and quality indicators, and determines the optimal process parameter combination through the optimization algorithm. The final output production control plan contains complete process parameter setting values ​​and control strategies.

[0154] For example, a batch of graphite composite insulation boards are evaluated for quality. In the thermal conductivity test, the measurement point spacing is set to 50mm to form a 16×16 measurement point matrix. The thermal conductivity of each measurement point is tested using the transient hot wire method, the heating time is 60s, and the temperature response curve is recorded. For the density test, samples are uniformly taken on the product, and each sample is measured 5 times to take the average value. For the strength test, specimens are prepared according to standard dimensions and compressive strength tests are performed. The CT scan uses a spiral scanning mode with a tube voltage of 150kV and a tube current of 200μA to obtain high-resolution three-dimensional structural images. Internal defects, including bubbles, cracks, and delamination, are identified through image processing. Defect characteristics include parameters such as quantity, size, distribution, and shape. These defect characteristics are correlated with physical property parameters to establish a quality feature matrix.

[0155] Multi-scale decomposition uses discrete wavelet transform to decompose the evaluation parameters into four scales. Different scales reflect different characteristics of performance changes and help to discover local anomalies. The data fusion process uses evidence theory methods and combines expert knowledge to assign credibility to data from different sources. The fusion results form a three-dimensional quality evaluation surface, which intuitively displays the spatial distribution characteristics of product quality. The quality evaluation results are used to guide the optimization of process parameters. By analyzing the quality evaluation surface, the weak links in quality are identified and the process parameters are adjusted in a targeted manner. The optimization of process parameters uses genetic algorithms to find the optimal parameter combination while meeting multiple quality goals. The optimized process parameters have been verified in production, and the product quality has been significantly improved.

[0156] In a specific embodiment, in step S604, the process of performing data fusion calculation on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters may specifically include the following steps:

[0157] S6041. Normalize and calculate the standard deviation of the physical performance data in the partition evaluation parameters, and obtain the physical property fusion matrix through weight allocation and data dimensionality reduction analysis;

[0158] Specifically, the physical performance data in the partition evaluation parameters are normalized and the standard deviation is calculated, and the physical property fusion matrix is ​​obtained through weight assignment and data dimension reduction analysis. In the production process of graphite composite insulation board, the physical performance data may include the density, thermal conductivity, specific heat capacity, etc. of the board. The dimensions and numerical ranges of these physical performance parameters vary greatly, and they need to be normalized to be unified between 0-1 to eliminate the influence of the dimensions and numerical ranges. Normalization can use the maximum-minimum normalization method. Then, the standard deviation of the normalized data is calculated to obtain the discrete degree of each parameter. Next, according to expert experience or data analysis results, a weight is assigned to each physical performance parameter to indicate its degree of influence on the quality of the board. Finally, data dimension reduction methods such as principal component analysis (PCA) are used to compress multiple physical performance parameters into a few comprehensive indicators to form a physical property fusion matrix. For example, for a batch of graphite composite insulation boards, the normalized values ​​of the density, thermal conductivity, and specific heat capacity were measured to be 0.8, 0.6, and 0.9, respectively, with standard deviations of 0.1, 0.2, and 0.05, and weights of 0.4, 0.3, and 0.3, respectively. Through PCA dimensionality reduction, two principal components were obtained to form a 2×1 physical property fusion matrix.

[0159] S6042. Perform cross validation and correlation analysis on the physical property fusion matrix and mechanical property data to generate a mechanical feature sequence; the mechanical property data may include the compressive strength, flexural strength, shear strength, etc. of the plate. Match each row of the physical property fusion matrix with the mechanical property data of the corresponding batch of plates to form several physical property-mechanical property data pairs. Then, a cross validation method, such as k-fold cross validation, is used to randomly divide the data pairs into k subsets, and one of the subsets is selected as the test set each time, and the remaining subsets are used as the training set to establish a mapping relationship between the physical property fusion matrix and the mechanical properties, and evaluate the accuracy of the mapping on the test set. Through multiple cross validations, the average mapping error between the physical property fusion matrix and the mechanical properties is obtained. At the same time, the correlation coefficient between each column of the physical property fusion matrix and the mechanical property data is calculated, and the physical property indicators that are highly correlated with the mechanical properties are selected. Finally, the selected physical property indicators are sorted from high to low according to the correlation with the mechanical properties to form a mechanical feature sequence. For example, the correlation coefficients of the two principal components of the physical property fusion matrix with the compressive strength are 0.8 and 0.6, respectively, and the correlation coefficients with the flexural strength are 0.7 and 0.9, respectively. Then the mechanical characteristic sequence is (principal component 2, principal component 1).

[0160] S6043. The mechanical feature sequence and the structural feature data are reconstructed according to 8192 sampling points to obtain feature fusion parameters; the structural feature data may include the size, thickness, honeycomb structure parameters, etc. of the plate. Since the number of sampling points of the structural feature data may be much larger than the length of the mechanical feature sequence, it is necessary to interpolate the mechanical feature sequence to make its length consistent with the number of sampling points of the structural feature data. Common interpolation methods include linear interpolation and spline interpolation. After interpolation, the mechanical feature sequence and the structural feature data are spliced ​​by column to form a feature fusion matrix. Next, the feature fusion matrix is ​​sampled, and 8192 sampling points are selected to obtain feature fusion parameters. The sampling points can be selected by equal interval sampling, random sampling, etc. For example, after the mechanical feature sequence is interpolated, the length is 10,000, and the number of sampling points of the structural feature data is 10,000. After splicing, a 10,000×8 feature fusion matrix is ​​formed. 8192 points are sampled at equal intervals from the matrix to obtain 8192×8 feature fusion parameters.

[0161] S6044. Perform principal component extraction and singular value decomposition calculation on the feature fusion parameters to form a feature mapping matrix; principal component extraction is a commonly used data dimensionality reduction method that can compress high-dimensional data into a low-dimensional space and extract the main features of the data. Singular value decomposition can decompose a matrix into the product of a left singular matrix, a singular value matrix, and a right singular matrix, where the column vectors of the left and right singular matrices represent the principal components of the original matrix. Perform principal component extraction on the feature fusion parameters to obtain a principal component matrix and corresponding singular values. Select the left and right singular vectors corresponding to the first few largest singular values ​​to form the basis of the row space and column space of the feature mapping matrix. For example, perform principal component extraction on the 8192×8 feature fusion parameters to obtain an 8×8 principal component matrix and an 8×1 singular value vector. Select the left singular vectors corresponding to the first three largest singular values ​​to form a 3×8 row space basis matrix, select the right singular vectors corresponding to the first three largest singular values ​​to form an 8×3 column space basis matrix, and multiply the two to obtain a 3×3 feature mapping matrix.

[0162] S6045. Use the feature mapping matrix to perform multidimensional data space transformation and coordinate mapping processing to obtain surface construction parameters; regard the feature fusion parameters as a set of vectors in the high-dimensional space, and use the feature mapping matrix to map them to the low-dimensional space. Specifically, the feature fusion parameter matrix is ​​multiplied by the column space basis matrix of the feature mapping matrix to obtain the coordinate matrix in the low-dimensional space. Then, according to the coordinate value of each vector in the coordinate matrix, a scatter plot is constructed in the low-dimensional space. Next, a surface fitting method such as the least squares method, spline fitting, etc. is used to fit a smooth surface on the scatter plot to obtain the surface construction parameters. The surface construction parameters can be the coefficients of the surface equation, the coordinates of the control points, etc. For example, the 8192×8 feature fusion parameter matrix is ​​multiplied by the 8×3 column space basis matrix to obtain the 8192×3 coordinate matrix. 8192 scattered points are drawn in the three-dimensional space, and then a quadratic surface is fitted by the least squares method to obtain the coefficients of the surface equation as the surface construction parameters.

[0163] S6046. Use the surface construction parameters to perform isosurface fitting and boundary continuity constraint analysis, and output the quality evaluation surface. According to the surface construction parameters, an initial surface is generated in three-dimensional space. Then, a series of isovalue points are selected on the initial surface, and the surface function values ​​at these points are equal. Use the isovalue points to fit a series of isovalue lines, and then fit the isovalue surface based on the isovalue lines. During the fitting process, it is necessary to impose continuity constraints on the boundaries of the isovalue surface so that it can smoothly transition with the adjacent isovalue surfaces at the boundaries to avoid sharp edges or discontinuities. Finally, the fitted isovalue surface is output as a quality evaluation surface. The quality evaluation surface intuitively shows the distribution of various performance indicators of the graphite composite insulation board. The higher the performance indicator value, the higher the surface. By observing the shape and trend of the quality evaluation surface, the overall quality level and weak links of the board can be quickly judged, providing a basis for the optimization control of the production process.

[0164] For example, there are 100 graphite composite insulation boards in a batch. Five physical performance parameters and three mechanical performance parameters are measured for each board. At the same time, three structural characteristic parameters of the board, namely, length, width, and thickness, are recorded. First, the five physical performance parameters are normalized and the standard deviation is calculated to obtain a 100×5 normalized matrix and a 1×5 standard deviation vector. After weight allocation and PCA dimension reduction, a 100×2 physical property fusion matrix is ​​obtained. Then, the physical property fusion matrix is ​​divided into a training set of 80 boards and a test set of 20 boards. The mapping relationship from the physical property fusion matrix to the mechanical properties is established on the training set, and cross-validation is performed on the test set, and the average prediction error is 5%. At the same time, the correlation coefficient between each column of the physical property fusion matrix and the three mechanical performance parameters is calculated. It is found that the first principal component has the highest correlation coefficient with the compressive strength, which is 0.9, and the second principal component has the highest correlation coefficient with the flexural strength, which is 0.8. Based on this, a mechanical feature sequence (first principal component, second principal component) is generated. Next, the mechanical feature sequence is interpolated to make its length consistent with the number of sampling points of the structural feature data (100×10=1000). The interpolated mechanical feature sequence and the three structural feature parameters are spliced ​​by column to obtain a 1000×5 feature fusion matrix. 8192 points are sampled at equal intervals from the matrix to obtain 8192×5 feature fusion parameters. Principal component extraction and singular value decomposition are performed on the feature fusion parameters to obtain a 5×5 principal component matrix and a 5×1 singular value vector. The left and right singular vectors corresponding to the first three largest singular values ​​are selected to form a 3×5 row space basis matrix and a 5×3 column space basis matrix, and the two are multiplied to obtain a 3×3 feature mapping matrix. The 8192×5 feature fusion parameter matrix is ​​multiplied with the 5×3 column space basis matrix to obtain an 8192×3 coordinate matrix. 8192 scattered points are plotted in three-dimensional space, and a quadratic surface is fitted using the least squares method. The coefficients of the surface equation are obtained as the surface construction parameters. Finally, 100 isolines were generated on the fitted quadratic surface, and then 100 isosurfaces were fitted to form the quality evaluation surface. By observing the quality evaluation surface, it was found that the surface was convex as a whole, but there were a few concave areas at the edges, indicating that the overall quality of this batch of insulation boards was good, but some individual indicators needed to be further improved.

[0165] The above describes the graphite composite insulation board production process control method in the embodiment of the present application. The following describes the graphite composite insulation board production process control system in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the control system for the production process of the graphite composite insulation board includes:

[0166] An analysis module is used to analyze and process the raw materials to obtain raw material characteristic data, and optimize the proportioning parameters based on the raw material characteristic data to obtain pre-treated raw materials;

[0167] A monitoring module, used to monitor and process the state parameters corresponding to the pre-treated raw material to obtain reaction state data, and use the reaction state data to control the process parameters to form a modified raw material;

[0168] An optimization module, used to perform a formulation optimization process according to the material parameters of the modified raw material, obtain formulation data, complete a material performance prediction process according to the formulation data, and generate an optimized formulation material;

[0169] A control module, used for performing molding analysis processing through the physical parameters of the optimized formula material to obtain molding process parameters, and implementing process control processing based on the molding process parameters to produce molding materials;

[0170] A processing module, used for performing curing analysis processing using the state parameters of the molding material to obtain curing process parameters, and performing curing process control processing according to the curing process parameters to form a cured product;

[0171] A generation module is used to perform quality evaluation processing based on the inspection data of the cured product, obtain evaluation parameters, execute control strategy generation processing according to the evaluation parameters, and generate a production control plan.

[0172] Through the coordinated cooperation of the above-mentioned components, the raw materials are analyzed and processed to achieve accurate acquisition of raw material characteristics and optimization of ratio parameters, ensure the quality stability of pre-treated raw materials, and provide a reliable raw material basis for subsequent processing; according to the state parameters corresponding to the pre-treated raw materials, monitoring and processing are carried out, and a real-time monitoring mechanism for reaction state data is established. Through the dynamic regulation of process parameters, the controllability of the modification process is ensured, and the performance consistency of the modified raw materials is improved; the formula optimization processing is carried out according to the material parameters of the modified raw materials, and the scientific management of the formula data is realized. Through the material performance prediction processing, the performance predictability of the optimized formula materials is guaranteed; by optimizing the physical parameters of the formula materials, the molding analysis processing is carried out, and the precise control system of the molding process parameters is established. Based on the process regulation of the molding process parameters, the quality level of the molding materials is improved; the state parameters of the molding materials are used for curing analysis processing, and the intelligent control of the curing process is realized. Through the precise regulation of the curing process, the performance stability of the cured products is guaranteed; based on the detection data of the cured products, quality evaluation processing is carried out, and a complete quality evaluation system is established. Through the optimization and generation of control strategies, closed-loop control of the production process is achieved, and the consistency of product quality and production efficiency are significantly improved. The present invention realizes intelligent control of the entire process from raw materials to finished products by establishing a multi-dimensional data collection, processing and control system, effectively solving the problems existing in traditional production methods such as large quality fluctuations, insufficient optimization of process parameters, and low process control accuracy.

[0173] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer executes the steps of the graphite composite insulation board production process control method.

[0174] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0175] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A graphite composite insulation board production process control method, characterized in that: The steps include: S1, raw material analysis stage: analyzing and processing the raw materials to obtain raw material characteristic data, optimizing the proportioning parameters based on the raw material characteristic data to obtain pre-treated raw materials; S2, monitoring and processing stage: monitoring and processing are performed according to the state parameters corresponding to the pre-treated raw materials to obtain reaction state data, and the process parameters are adjusted and processed using the reaction state data to form modified raw materials; S3, formulation optimization stage: performing formulation optimization processing according to the material parameters of the modified raw material to obtain formulation data, completing material performance prediction processing according to the formulation data, and generating optimized formulation materials; S4, molding analysis stage: performing molding analysis processing through the physical parameters of the optimized formula material to obtain molding process parameters, and implementing process control processing based on the molding process parameters to produce molding materials; S5, curing analysis stage: using the state parameters of the molding material to perform curing analysis processing to obtain curing process parameters, and implementing curing process control processing according to the curing process parameters to form a cured product; S6, quality evaluation stage: performing quality evaluation processing based on the inspection data of the cured product to obtain evaluation parameters, executing control strategy generation processing according to the evaluation parameters to generate a production control plan.

2. The graphite composite insulation board production process control method according to claim 1, characterized in that: The raw material analysis stage of step S1 includes the following steps: S101, collecting spectra of the raw material in multiple time periods, and obtaining spectral characteristic data of the raw material by comparing the spectral intensity with the standard spectrum; S102, calculating the time distribution of the physicochemical characteristics, particle size distribution and morphological characteristics according to the raw material spectral characteristic data, and generating initial characteristic standard data; S103, performing partition calculation on the initial characteristic standard data according to 64 data channels to obtain partition characteristic data; S104, performing cumulative calculation on the physical features, chemical features and morphological features in the partition feature data to form a feature cumulative spectrum; S105, using the characteristic cumulative spectrum to perform distribution calculation on each data channel of the raw material to obtain raw material characteristic data; S106, performing correlation analysis on the raw material characteristic data and the ratio optimization parameters, and outputting the pre-processed raw material.

3. The graphite composite insulation board production process control method according to claim 1, characterized in that: The monitoring and processing stage of step S2 includes the following steps: S201. The pretreated raw material is subjected to temperature field acquisition in multiple time periods, and temperature change data is obtained by comparative analysis of the temperature field intensity and the standard temperature field; S202. Calculate the time distribution characteristics of the reaction rate, conversion rate and dispersion degree according to the temperature change data to generate reaction state data; S203. The reaction state data is divided into regions according to 256 data sampling points to obtain partitioned reaction state data; S204. Perform weighted cumulative calculation on the reaction rate, conversion rate and dispersion degree in the partition reaction state data to form a state cumulative distribution; S205. Using the state cumulative distribution to perform distribution fitting calculation on the state parameters of each sampling point to obtain reaction state data; S206. Compare and analyze the reaction state data with the process parameter threshold range, and output the modified raw material.

4. The graphite composite insulation board production process control method according to claim 1, characterized in that: The recipe optimization stage of step S3 includes the following steps: S301. Collecting the physical property parameters of the modified raw material in multiple time periods, and obtaining a material parameter sequence by comparative analysis of thermal conductivity, density and strength; S302. Calculate the time variation characteristics of material strength, thermal conductivity and density distribution according to the material parameter sequence to generate formulation data; S303. The formula data is numerically decomposed and calculated according to 512 sampling points to obtain partition formula data; S304. Perform weighted statistical calculations on the physical properties, mechanical properties and thermal conductivity in the partitioned formula data to form a formula distribution spectrum; S305. Perform data regression analysis on each sampling point using the recipe distribution spectrum to obtain recipe data; S306. Match and analyze the formula data with the performance index range, and output optimized formula materials.

5. The method for controlling the production process of graphite composite insulation board according to claim 1, characterized in that: The forming analysis phase of step S4 includes the following steps: S401. Collecting the temperature field and pressure field of the optimized formula material at multiple points, and obtaining the molding process data by comparative analysis of the temperature field and pressure field gradients; S402. Calculate the time evolution characteristics of material stress distribution, deformation and density distribution according to the molding process data to generate molding process parameters; S403. The forming process parameters are numerically discretely calculated according to 1024 grid nodes to obtain partition forming parameters; S404. Perform multi-field coupling calculation on the temperature field, pressure field and deformation field in the partition molding parameters to form a molding state curve; S405. Perform trend analysis on each monitoring point using the molding state curve to obtain molding process parameters; S406: Perform deviation analysis on the molding process parameters and process thresholds, and output molding materials.

6. The graphite composite insulation board production process control method according to claim 1, characterized in that: The curing analysis stage of step S5 includes the following steps: S501. Performing time series acquisition of the temperature field and infrared radiation intensity monitoring at multiple measuring points on the molding material, and obtaining a curing state sequence through spatiotemporal evolution analysis of the temperature field distribution and the radiation intensity gradient; S502. Calculate the dynamic change characteristics of the crosslinking degree distribution, the curing depth and the stress distribution according to the curing state sequence, and generate curing process parameters; S503. The curing process parameters are classified and processed according to 2048 sampling points to obtain partition curing parameters; S504. Perform multi-field collaborative calculation on the temperature field evolution, stress distribution and curing degree in the partition curing parameters to form a curing dynamic spectrum; S505. Using the curing dynamic spectrum, data reconstruction and analysis are performed on the curing degree of each sensor node to obtain curing process parameters; S506. Perform process trajectory planning and curve fitting analysis on the curing process parameters, and output a cured product.

7. The graphite composite insulation board production process control method according to claim 1, characterized in that: The quality evaluation stage in step S6 includes the following steps: S601. Perform multi-dimensional physical property index detection and internal structure scanning on the cured product, and obtain a quality feature matrix through correlation analysis between physical property parameter distribution and internal defect characteristics; S602. Calculate the spatial distribution law of intensity distribution, thermal conductivity and density uniformity according to the quality characteristic matrix, and generate evaluation parameters; S603. Perform multi-scale decomposition processing on the evaluation parameters according to 4096 feature points to obtain partition evaluation parameters; S604. Perform data fusion calculation on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters to form a quality evaluation surface; S605. Perform sensitivity analysis on the abnormality of each test item using the quality evaluation surface to obtain evaluation parameters; S606. Perform multi-objective optimization and parameter mapping analysis on the evaluation parameters and process control variables, and output a production control plan.

8. The method for controlling the production process of graphite composite insulation board according to claim 7, characterized in that: In step S604, data fusion calculation is performed on the physical properties, mechanical properties and structural characteristics in the partition evaluation parameters to form a quality evaluation surface, including the following steps: S6041. Normalizing and calculating the standard deviation of the physical performance data in the partition evaluation parameters, and obtaining a physical property fusion matrix through weight allocation and data dimensionality reduction analysis; S6042. Perform cross validation and correlation analysis based on the physical property fusion matrix and mechanical performance data to generate a mechanical feature sequence; S6043. Reconstruct the mechanical feature sequence and the structural feature data according to 8192 sampling points to obtain feature fusion parameters; S6044. Perform principal component extraction and singular value decomposition calculation on the feature fusion parameters to form a feature mapping matrix; S6045. Perform multidimensional data space transformation and coordinate mapping processing using the feature mapping matrix to obtain surface construction parameters; S6046. Perform isosurface fitting and boundary continuity constraint analysis using the surface construction parameters, and output a quality evaluation surface.

9. A graphite composite insulation board production process control system, used to implement the graphite composite insulation board production process control method according to any one of claims 1 to 8, characterized in that: The graphite composite insulation board production process control system includes: An analysis module is used to analyze and process the raw materials to obtain raw material characteristic data, and optimize the ratio parameters based on the raw material characteristic data to obtain pre-processed raw materials; A monitoring module, used to monitor and process the state parameters corresponding to the pre-treated raw material to obtain reaction state data, and use the reaction state data to control the process parameters to form a modified raw material; An optimization module, used to perform a formulation optimization process according to the material parameters of the modified raw material, obtain formulation data, complete a material performance prediction process according to the formulation data, and generate an optimized formulation material; A control module, used for performing molding analysis processing through the physical parameters of the optimized formula material to obtain molding process parameters, and implementing process control processing based on the molding process parameters to produce molding materials; A processing module, used for performing curing analysis processing using the state parameters of the molding material to obtain curing process parameters, and performing curing process control processing according to the curing process parameters to form a cured product; A generation module is used to perform quality evaluation processing based on the inspection data of the cured product, obtain evaluation parameters, execute control strategy generation processing according to the evaluation parameters, and generate a production control plan.

10. A storage medium, which is a computer-readable storage medium storing instructions, characterized in that: When the instructions are executed by the processor, the graphite composite insulation board production process control method according to any one of claims 1 to 8 is implemented.

Images