Digital visual control system and method for synthetic mica material production line

Abstract

The invention relates to the technical field of industrial production line control, in particular to a digital visual control system and method for a synthetic mica material production line. The method comprises the following steps: acquiring industrial data of a mica material, and extracting a data feature vector to obtain a batch quality coefficient, temperature-pressure fitting data and lattice distortion data; outputting a mica material defect distribution diagram by taking the temperature-pressure fitting data unfolded along the direction of the kiln as an abscissa, taking the lattice distortion data as an axis coordinate and combining with the batch quality coefficient; therefore, the digital visual control system based on multi-dimensional data fusion is constructed, the key technical problems that in the traditional mica material production process, technological parameters are difficult to control accurately, product quality fluctuation is large, and the defect rate is high are solved, and the production efficiency, the product quality stability and the technological control intelligent level are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial production line control, and in particular to a digital visual control system and method for a synthetic mica material production line. BACKGROUND

[0002] The existing method can only independently monitor the kiln temperature, pressure and crystal growth parameters (such as thermocouple gradient, pressure spectrum), but lacks correlation modeling of thermal data (axial temperature gradient, pressure pulsation spectrum) and material intrinsic properties (lattice distortion, birefringence). Especially, it is impossible to construct a defect distribution model through the spatial mapping relationship of temperature-pressure fitting data and lattice distortion, making it difficult to locate key problems such as cold zone density anomaly (> 15% threshold) and pinhole defects (probability > 30%). The traditional method needs to manually preset a static density difference threshold (such as a crystal nucleus density deviation > 15%) and a defect probability threshold (such as a pinhole defect > 30%), but the dynamic interference in actual production (such as raw material fluorine content fluctuation, kiln thermal inertia) leads to poor adaptability of the fixed threshold. The identification of suspended crystal nuclei in the crystal phase region depends on manual interpretation, and the melt viscosity distribution map (based on fluorine content labeled crystal nucleus data) and the quantitative correlation rule of production line defects (such as triggering defect labeling when the reaction rate < 0.8 and the viscosity standard deviation > 50 Pa·s) are not established. SUMMARY

[0003] Therefore, it is necessary to provide a digital visual control system and method for a synthetic mica material production line to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a digital visual control method for a synthetic mica material production line, the method comprising the following steps:

[0005] Step S1: Obtain mica material industrial data and extract data feature vectors to obtain batch quality coefficients, temperature-pressure fitting data and lattice distortion data;

[0006] Step S2: Expand the temperature-pressure fitting data along the kiln direction as the horizontal coordinate, and take the lattice distortion data as the axis coordinate, and output the mica material defect distribution map combined with the batch quality coefficients;

[0007] Step S3: Determine the temperature cold zone according to the mica material defect distribution map, if the density difference of the temperature cold zone is greater than the preset density difference threshold, then perform axial temperature compensation to obtain a temperature compensation region; and according to the mica material defect distribution map, if the pinhole defect probability is greater than the preset defect probability threshold, then perform fluorine compensation to obtain a fluorine compensation region;

[0008] Step S4: obtaining an actual mica crystal growth image; superimposing the temperature compensation region and the fluorine compensation region according to the current timestamp, and performing similarity matching with the actual mica crystal growth image, so as to display a mica material whole-process visualization graph and generate a mica material whole-process report;

[0009] Step S5: generating a control instruction of the mica material production line based on the mica material whole-process report, and performing whole-process control on the mica material production line according to the control instruction.

[0010] In the present specification, an information interaction control system of a vital sign monitor is provided for executing the information interaction control method based on the vital sign monitor described above, and the information interaction control system of the vital sign monitor comprises:

[0011] A data acquisition and feature construction module is configured to acquire mica material industrial data, extract a data feature vector, obtain a batch quality coefficient, temperature-pressure fitting data, and lattice distortion data.

[0012] A defect modeling and probability analysis module is configured to expand the temperature-pressure fitting data along the kiln direction as the horizontal coordinate, take the lattice distortion data as the axis coordinate, and output a mica material defect distribution graph in combination with the batch quality coefficient.

[0013] A defect compensation strategy generation module is configured to determine a temperature cold zone according to the mica material defect distribution graph, perform axial temperature compensation to obtain a temperature compensation region if the density difference of the temperature cold zone is greater than a preset density difference threshold, and statistically analyze the pinhole probability according to the mica material defect distribution graph, and perform fluorine compensation to obtain a fluorine compensation region if the pinhole defect probability is greater than a preset defect probability threshold.

[0014] A visualization and report generation module is configured to obtain an actual mica crystal growth image; superimpose the temperature compensation region and the fluorine compensation region according to the current timestamp, and perform similarity matching with the actual mica crystal growth image, so as to display a mica material whole-process visualization graph and generate a mica material whole-process report.

[0015] A control instruction module is configured to generate a control instruction of the mica material production line based on the mica material whole-process report, and perform whole-process control on the mica material production line according to the control instruction.

[0016] The beneficial effects of the present application are: by collecting the axial temperature gradient data (range 0.88-0.95 ℃ / cm) and pressure pulsation frequency spectrum data (range 0.32-0.41 kPa) of the six temperature zones of the kiln in real time, a temperature-pressure fitting model is established, and the temperature cold zone phenomenon existing in the third and fourth temperature zones of the middle section of the kiln is effectively identified, the crystal nucleus density deviation reaches negative fifteen percent, after the temperature in this region is raised by 2.3 degrees Celsius through the axial temperature compensation strategy, the crystal nucleus density uniformity is improved by twenty-two percent, and the product quality coefficient is improved from 0.82 to 0.86.

[0017] Statistical analysis of twenty production batches shows that the mica crystal nucleus density and the batch quality coefficient show a significant positive correlation, when the crystal nucleus density is maintained in the optimized range of 90-100 per meter, the batch quality coefficient is stable above 0.85, and the product qualified rate is above ninety-five percent. Defect probability analysis data show that the pinhole defect probability of temperature zone four under the traditional production mode is as high as thirty-five percent, after the fluorine content is accurately adjusted by 0.024 percent through the fluorine compensation strategy, the pinhole defect probability is reduced to twelve percent, the microcrack defect probability is reduced from twenty-eight percent to eight percent, and the overall defect rate is reduced from thirty-one percent to eighteen percent.

[0018] Mica melt viscosity distribution data show that the viscosity can be accurately controlled along the axial direction of the kiln, the viscosity standard deviation is reduced from 75 Pa·s of the traditional process to 45 Pa·s, the viscosity distribution uniformity is improved by forty percent, thereby ensuring the stability and consistency of the crystal growth process, and finally realizing the significant improvement of the production line automation control precision and the stable optimization of the product quality. Therefore, by constructing a multi-dimensional data fusion digital visual control system, the present application solves the key technical problems of difficult accurate control of process parameters, large product quality fluctuation and high defect rate in the traditional mica material production process, improves the production efficiency, product quality stability and process control intelligent level. BRIEF DESCRIPTION OF DRAWINGS

[0019] Fig. 1 It is a step flowchart of a digital visual control method for a synthetic mica material production line;

[0020] Fig. 2 It is a schematic diagram of the axial temperature gradient distribution of the kiln;

[0021] The implementation, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0022] The technical method of the patent of the present application will be described clearly and completely in combination with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0023] In addition, the drawings are only schematic illustrations of the present application and are not necessarily drawn to scale. Identical reference signs refer to identical or similar parts throughout the drawings and thus repeated descriptions will be omitted. Some of the block diagrams shown in the drawings are functional entities, which do not necessarily have to correspond to physically or logically independent entities. The functional entities can be implemented in software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0024] It should be understood that although the terms "first", "second" and the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the example embodiments, a first element can be called a second element, and similarly a second element can be called a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] To achieve the above-mentioned purpose, please refer to Figs. 1-2 A digital visualization control method for a synthetic mica material production line, the method comprising the following steps:

[0026] In the embodiments of the present application, referring to Fig. 1 The step flow diagram of the digital visualization control method for the synthetic mica material production line of the present application is shown in the present example. In the present example, the digital visualization control method for the synthetic mica material production line comprises the following steps:

[0027] Step S1: Obtain mica material industrial data and extract a data feature vector to obtain a batch quality coefficient, temperature-pressure fitting data and lattice distortion data;

[0028] Preferably, step S1 comprises the following steps:

[0029] Obtain mica material industrial data, wherein the mica material industrial data includes a raw material feature vector, thermal tensor data and crystal growth intrinsic quantity; the thermal tensor data includes an axial temperature gradient of six temperature zones of thermocouples and a pressure pulsation frequency spectrum of a pressure transmitter;

[0030] Batch quality is encoded with raw material characteristic vector based on fluorine content to obtain batch quality coefficient;

[0031] Wavelet fusion is performed on axial temperature gradient of thermal tensor data and pressure pulsation spectrum to output temperature-pressure fitting data;

[0032] Biaxial interference of crystal growth intrinsic quantity is analyzed to obtain crystal lattice distortion data.

[0033] In the embodiment of the present application, the raw material characteristic vector is obtained, which contains the percentage content data of silicon dioxide, aluminum oxide, potassium oxide and fluorine element, forming a four-dimensional chemical composition vector; the thermal tensor data is collected, specifically including the axial temperature gradient values measured by six temperature zone thermocouple sensors arranged along the longitudinal axis of the kiln and the pressure pulsation spectrum information detected by the pressure transmitter; in addition, crystal growth intrinsic quantity data is also needed to be obtained, mainly the crystal lattice structure parameters obtained by X-ray diffraction detection.

[0034] In the data processing link, batch quality encoding algorithm is established based on fluorine content as the reference value, and batch quality coefficient is obtained by calculating the weighted deviation of each chemical component relative to the standard ratio; for thermal data processing, wavelet fusion technology is adopted to perform multi-scale decomposition and reconstruction on the axial temperature gradient data of the six temperature zones and the corresponding pressure pulsation spectrum main frequency amplitude, to generate a fitting parameter set describing the temperature-pressure coupling relationship; in the aspect of crystal lattice distortion analysis, through analyzing the biaxial interference phenomenon in the crystal growth process, the extraordinary refractive index and ordinary refractive index data are extracted, the birefringence compensation amount is calculated combined with the crystal thickness information, and then the quantitative crystal lattice distortion characteristic data is obtained through optical path simulation.

[0035] In one implementation manner of the embodiment of the present application, it is assumed that the raw material characteristic vector shows that the silicon dioxide content is 45.2%, the aluminum oxide content is 35.8%, the potassium oxide content is 10.5%, and the fluorine content is 4.3%, and the batch quality coefficient is calculated to be 0.83 based on the standard fluorine content of 4.5%;

[0036] The axial temperature gradients measured by the six temperature zone thermocouples are 0.88, 0.92, 0.95, 0.89, 0.85 and 0.91 respectively;

[0037] The corresponding pressure pulsation spectrum integral values are 0.32, 0.38, 0.41, 0.35, 0.33 and 0.39 kilopascals per centimeter per degree Celsius.

[0038] After wavelet fusion processing, the output temperature-pressure fitting data is a five-order polynomial coefficient set; at the same time, the X-ray diffraction detection shows that the crystal thickness is 120 microns, the extraordinary refractive index is 1.6012, and the ordinary refractive index is 1.5935, and the birefringence compensation amount formula Δδ=(n e -no The compensation amount was calculated to be 0.924 micrometers by d-δ, and the lattice distortion data for this batch was finally obtained as 0.25% by optical path difference simulation.

[0039] Preferably, the biaxial interference of wavelet fusion and analysis of crystal growth eigenvalues ​​based on the axial temperature gradient and pressure pulsation spectrum of the thermal tensor data includes:

[0040] Six temperature zone acquisition points were set along the longitudinal axis of the kiln, and envelope extraction was performed on the six temperature zone acquisition points to obtain the axial temperature gradient; the pressure spectrum energy of the kiln was integrated by signal amplitude to obtain the pressure pulsation spectrum;

[0041] The axial temperature gradient and pressure pulsation spectrum are fused using wavelet fusing and sorted according to temperature-pressure to obtain temperature-pressure fitting data;

[0042] The thickness of the mica crystal was obtained; the extraordinary and ordinary refractive indices were extracted from the intrinsic properties of crystal growth, and the birefringence compensation data were calculated using the biaxial interference analytical formula, which is as follows:

[0043] Δδ=(n e -n o )d-δ;

[0044] Where Δδ is the birefringence compensation data, n e For the refractive index of light, n o denoted as the ordinary refractive index, d as the mica crystal thickness, and δ as the mica birefringence.

[0045] The optical path is simulated using birefringence compensation data, and the lattice distortion data is obtained by compensating for the difference based on the simulated optical path.

[0046] Please see Fig. 2 The red line represents the axial temperature gradient; the blue line represents the pressure pulsation spectrum; and the blue line shows the distribution of the axial temperature gradient and pressure pulsation spectrum collected from six temperature zones within the kiln.

[0047] In one implementation of this invention, wavelet fusion uses the db4 wavelet basis for 5-level decomposition, extracts low-frequency coefficients to reconstruct the fused signal, and then arranges the data points in ascending order of temperature value, using a cubic polynomial to generate temperature-pressure fitting parameters.

[0048] In one implementation of this invention, the strain conversion coefficient is determined through a material calibration experiment.

[0049] In one implementation of this invention, it is assumed that the temperature envelope derivative G T [3] = 0.92℃ / cm, pressure spectrum integral P f[3] = 0.38 kPa.

[0050] Wavelet fusion fitting coefficient B3 = [0.41, -1.7 x 10 -3 , 3.0 x 10 -6 , -6.9 x 10 -10 ].

[0051] Crystal thickness d = 120 μm, n e = 1.6012 at position (50, 50), n o = 1.5935, theoretical δ ref = 0.005,

[0052] Calculated Δδ = (1.6012 - 1.5935) x 120 - 0.005 = 0.924 μm.

[0053] Optical path difference simulation result ε RMS = 0.25%, corresponding to lattice distortion ε = 0.25% (k = 1).

[0054] Step S2: expand the temperature-pressure fitting data along the kiln direction as the horizontal coordinate, take the lattice distortion data as the axis coordinate, and output the mica material defect distribution map in combination with the batch quality coefficient;

[0055] Preferably, step S2 comprises the following steps:

[0056] Step S21: expand the temperature-pressure fitting data along the kiln direction as the horizontal coordinate, take the lattice distortion data as the axis coordinate, and construct a spatial mica evolution coordinate system;

[0057] Step S22: locate the crystal phase region of the spatial mica evolution coordinate system, identify the suspended mica crystal nucleus of the crystal phase region, generate mica crystal nucleus identification data, and mark the mica crystal nucleus identification data according to the fluorine content, thereby outputting a mica melt viscosity distribution map;

[0058] Step S23: find the line defects of the mica melt viscosity distribution map based on the batch quality coefficient, and obtain a mica material defect distribution map.

[0059] In the embodiment of the application, the temperature-pressure fitting data corresponding to each temperature zone (usually including polynomial coefficients or key fitting values) is taken as the representative value of the kiln axial position of the temperature zone (for example, the temperature zone 1 is at 0-5 meters, the temperature zone 2 is at 5-10 meters...), so as to establish the horizontal coordinate axis (X axis) covering the entire kiln length; at the same time, the lattice distortion data (such as average distortion rate or key point distortion variable) associated with the corresponding position or batch is taken as the vertical coordinate axis (Y axis), thereby constructing a two-dimensional coordinate system, each data point of which represents the correlation between the process state (temperature-pressure fitting result) and the material structure state (lattice distortion) of a certain axial position (or temperature zone) of the kiln.

[0060] In this coordinate system, a threshold range (e.g. Y value between 0.1% and 0.3%) is set based on the lattice distortion data (Y value) to frame specific "crystal phase regions" which represent the kiln positions where the crystal structure features are consistent with a specific growth stage; then, the original crystal image data or interference data corresponding to these crystal phase regions are processed using image analysis or pattern recognition algorithms (e.g. connected component analysis, feature point detection) to identify the tiny crystal nuclei presenting suspended independent growth features in the images, and output the position coordinates (X value in the coordinate system) and feature parameters (e.g. size, contour intensity) of each identified crystal nucleus as "mica crystal nucleus identification data"; next, according to the fluorine content (scalar value or spatial distribution value) in the production batch raw material data, these identified crystal nucleus data points are labeled with fluorine content labels (e.g. high fluorine area, low fluorine area); finally, based on the material rheology model, the viscosity estimation value of the mica melt at this position is calculated using the crystal nucleus distribution density (e.g. the number of crystal nuclei per unit axial length) and feature parameters with fluorine content labels, thus generating a distribution map showing the variation of melt viscosity at different positions along the kiln axis.

[0061] The batch quality coefficient (a comprehensive quality score Q, e.g. in the range of 0-1) is taken as a reference benchmark, and on the generated viscosity distribution map, those areas with viscosity values significantly deviating from the expected range (e.g. too high leading to poor flowability, or too low leading to instability) are located; these deviating areas are identified as potential "production line defect" positions, and combined with the original lattice distortion data (Y value) and crystal nucleus distribution information at these positions, a "mica material defect distribution map" integrating kiln position, process state (temperature-pressure fitting), material structure state (lattice distortion), melt state (viscosity anomaly), and defect marking is finally output.

[0062] In one implementation of an embodiment of the present application, a 30-meter kiln is assumed to be divided into 6 5-meter temperature zones (X axis: temperature zone 1 = 0-5m, temperature zone 2 = 5-10m,..., temperature zone 6 = 25-30m).

[0063] The temperature-pressure fitting value of temperature zone 4 (X = 15-20m) is 0.75 (representing a specific process state), and its corresponding lattice distortion data Y = 0.22%.

[0064] The area with lattice distortion Y value between 0.2% and 0.25% (including most of temperature zone 4 and parts of temperature zones 3 and 5) is identified as a key crystal phase region. In the crystal image of temperature zone 4 (around X = 17m), 5 suspended crystal nuclei (average diameter 10μm) are identified, and this area is labeled as "medium" (4.0%) in terms of fluorine content. According to the crystal nucleus density and fluorine content, the average melt viscosity of temperature zone 4 is calculated to be 1200Pa·s.

[0065] Batch quality factor Q = 0.85 (good). On the viscosity profile, it was found that the viscosity of zone 2 (X = 5-10 m) was as high as 1800 Pa·s (significantly higher than the normal range of 1200 ± 200 Pa·s), and the lattice distortion Y = 0.35% of this area was also high. Therefore, on the final defect profile, zone 2 was marked as a "high viscosity defect zone", and zone 4 was marked as a "normal zone".

[0066] Preferably, step S23 comprises the following steps:

[0067] Based on the mica melt viscosity profile, the reaction rate and viscosity standard deviation are extracted respectively to obtain the mica reaction rate data and viscosity standard deviation data;

[0068] Obtain a historical defect database; cross-correlate the mica melt viscosity profile based on the batch quality factor of the mica fluorine content, when the mica reaction rate data is less than 0.8 and the viscosity standard deviation data is greater than 50 Pa·s, use the historical defect database to mark production defects, and obtain mica production line marking data;

[0069] According to the mica production line marking data, the defect probability is calculated, when the single production line defect rate of the mica production line marking data is greater than 0.2, the mica material defect distribution map is constructed.

[0070] In the embodiment of the application, two key data are extracted from the generated mica melt viscosity profile (the profile takes the kiln axial position as the horizontal coordinate and the viscosity value of each position point as the vertical coordinate):

[0071] 1) Mica reaction rate data: calculate the rate of change of viscosity value along the kiln axial direction (production flow direction), that is, the average slope of viscosity change between adjacent position points (for example, position A viscosity 1200 Pa·s, position B (1 meter downstream) viscosity 1100 Pa·s, then the rate is -100 Pa·s / m);

[0072] 2) Viscosity standard deviation data: calculate the fluctuation size (dispersion degree) of the viscosity value of all position points in the entire kiln or a specific section (such as each temperature zone) relative to its average value, and obtain a statistical value (unit Pa·s). Then, obtain a historical defect database (which stores the actual defect type and position record of different fluorine content and different viscosity characteristics in past production batches).

[0073] Based on the fluorine content level implied by the batch quality factor of the current batch (for example, the quality factor Q = 0.85 corresponds to the fluorine content of about 4.2%), the historical records with similar fluorine content are screened out in the historical database. Cross-correlation processing is performed:

[0074] Compare the current batch's viscosity profile characteristics (especially the just-extracted reaction rate and viscosity standard deviation) with the filtered historical data.

[0075] Set a clear decision rule: if the current batch's reaction rate data (average change slope) is less than 0.8 Pa·s / m (meaning the viscosity change is too flat, and the reaction might not be sufficient) and the viscosity standard deviation data is greater than 50 Pa·s (meaning the viscosity fluctuates dramatically in space, and the process is unstable), then according to the high-frequency defects (such as "bubble aggregation", "grain coarsening") in the matching historical data with similar viscosity characteristics, mark these potential production defects on the current viscosity profile at the corresponding positions, and generate mica production line marking data containing defect types and positions (data structure: position coordinates + defect type label).

[0076] Finally, according to this marking data, calculate the proportion of the number of position points marked as defects to the total number of position points in the entire production line or key monitoring section, and obtain the single production line defect rate. If this defect rate is greater than 0.2 (i.e. 20% of the positions are marked as problematic), trigger the construction of a mica material defect distribution map (which integrates position, viscosity, marked defect type, and lattice distortion from previous steps, etc.).

[0077] In one implementation of an embodiment of the present application, assume that the current batch of mica melt viscosity profile shows that 60 points are measured along the 30-meter kiln (one point every 0.5 meters). The calculation results are:

[0078] Reaction rate data (average slope) = 0.65 Pa·s / m (less than the threshold value 0.8)

[0079] Viscosity standard deviation data = 55 Pa·s (greater than the threshold value 50 Pa·s)

[0080] The current batch quality coefficient Q = 0.86, and it is inferred that the fluorine content is about 4.3%. Querying the historical database finds that past batches with fluorine content between 4.2% and 4.4% and simultaneously meeting the "rate <0.8 & standard deviation >50" condition frequently appear "micro-crack" defects in the middle section of the kiln (10-20 meters).

[0081] Therefore, the system marks "micro-crack risk" on the data points in the 10-20 meter interval of the current viscosity profile, and generates production line marking data. Among the 60 points of the entire kiln, 15 points (concentrated in 10-20 meters) are marked, and the defect rate = 15 / 60 = 0.25 (greater than the threshold value 0.2). Therefore, a detailed mica material defect distribution map is constructed and output, which clearly indicates that 10-20 meters is a "high-risk area for micro-cracks", and superimposes the abnormal viscosity values and lattice distortion data in this area.

[0082] Step S3: judging temperature cold zone according to the mica material defect distribution map, if the density difference of the temperature cold zone is greater than a preset density difference threshold, then performing axial temperature compensation to obtain a temperature compensation region; and calculating the pinhole probability according to the mica material defect distribution map, if the pinhole defect probability is greater than a preset defect probability threshold, then performing fluorine compensation to obtain a fluorine compensation region.

[0083] Preferably, step S3 comprises the following steps:

[0084] calculating the deviation of the longitudinal mica crystal nucleus density according to the mica material defect distribution map, if the crystal nucleus density deviation is greater than a preset density difference threshold of 15%, then performing temperature compensation by using an axial temperature compensation formula to obtain a temperature compensation region;

[0085] extracting the current batch quality coefficient according to the mica material defect distribution map, if the pinhole defect probability of the current batch quality coefficient is greater than a preset defect probability threshold of 30%, then performing new fluorine content formula setting based on a new fluorine content formula setting formula to obtain a fluorine compensation region.

[0086] Preferably, the axial temperature compensation formula and the new fluorine content formula setting formula comprise:

[0087]

[0088] wherein ΔT is the axial temperature compensation, t1 and t2 are the start and end time of the integral time interval, f(Δρ) is the density deviation function, and Δρ is the mica crystal nucleus density difference.

[0089] ΔF = 0.2 x |K_batch 1 -1|;

[0090] wherein ΔF is the data of the new fluorine content formula setting, and K_batch 1 is the current batch quality coefficient.

[0091] In the embodiment of the application, based on the mica material defect distribution map (which contains information of different axial positions of the kiln, including lattice distortion, defect markers and key crystal nucleus density data), the actual mica crystal nucleus density (for example, the number of crystal nucleus per meter length) of each axial position (or small region) is calculated.

[0092] The average crystal nucleus density of the whole kiln is taken as a reference value, and the relative percentage deviation of the actual crystal nucleus density of each position from the average value is calculated (i.e. (local density-average density) / average density*100%). A preset density deviation threshold (15%) is set. If the absolute value of the calculated crystal nucleus density deviation of a certain position (or a continuous area) is greater than 15% (i.e. the crystal nucleus is significantly too much or too little), it is determined that there is a problem of uneven temperature distribution in the area. For these over-standard areas, the required temperature adjustment amount is calculated using the axial temperature compensation formula (the formula is Δρ is the calculated crystal nucleus density deviation value (for example, -20%) of the position, f(Δρ) is a function that determines the compensation strength according to the density deviation (the specific form is determined by the process, such as f(Δρ)=k*Δρ, k is a coefficient), t1 and t2 represent the time interval of the material in the kiln in this section.

[0093] The integral of this function in the residence time interval is obtained, and the final temperature compensation amount ΔT (unit: Celsius) is obtained. Thus, the temperature compensation area that needs to be adjusted in temperature is accurately defined (for example, the interval from X meters to Y meters of the kiln needs to be raised or lowered by ΔT). The quality coefficient K_batch 1 of the current batch is directly extracted from the batch information associated with the same defect distribution map (this is a comprehensive quality score, for example, ranging from 0.8 to 1.0).

[0094] According to the defect type data marked on the defect distribution map, the proportion of the position points of the pinhole defects in the total position points in the whole batch is calculated to obtain the pinhole defect probability of the current batch. A preset defect probability threshold (30%) is set. If the calculated pinhole defect probability is greater than 30%, it is determined that the fluorine content is the main cause.

[0095] The fluorine content adjustment amount is calculated using the new fluorine content formula setting formula (the formula is ΔF=0.2*|K_batch 1 -1|). Here, |K_batch 1 is the extracted quality coefficient of the current batch (for example, 0.82), and ΔF calculated by the formula is the amount (usually in percentage) that needs to be added to the original formula fluorine content. Thus, the fluorine compensation area that needs to adjust the raw material formula is determined (usually the raw material supply link of the whole batch needs to increase the fluorine content by ΔF.

[0096] In an implementation manner of the embodiment of the present application, the density deviation function f(Δρ) takes the residence time of the material through the kiln section (for example, 5 minutes).

[0097] In an implementation manner of the embodiment of the present application, it is assumed that the analysis on the defect distribution map obtains:

[0098] The crystal nucleus density of the kiln in a certain area (position 10-12 meters) is 85 / m, and the average crystal nucleus density of the whole kiln is 100 / m. The crystal nucleus density deviation Δρ = (85-100) / 100x100% = -15%. The absolute value |Δρ| = 15%, which is equal to the threshold value 15% (generally greater than or equal to the threshold value, which is triggered, assuming that it is triggered here). Assuming that the material residence time t2-t1 of the area is 5 minutes, f(Δρ) is defined as f(Δρ) = 0.1xΔρ (i.e. -0.1℃ per minute needs to be compensated for every 1% negative deviation).

[0099] Temperature compensation amount Therefore, the temperature compensation area is determined to be the 10-12 meter interval of the kiln, and a temperature drop of 7.5℃ is required.

[0100] The current batch quality coefficient K_batch1 = 0.82. The defect distribution chart statistics show that among the 60 monitoring points of the whole kiln, 25 points are marked as "pinhole defects". The pinhole defect probability = 25 / 60 ≈ 41.7% (greater than the threshold value 30%).

[0101] The fluorine content formula setting formula is applied: ΔF = 0.2x|0.82-1| = 0.2x0.18 = +0.036%.

[0102] Therefore, the fluorine compensation area indicates that the fluorine content of the whole batch of raw material formula needs to be increased by 0.036% (for example, from 4.500% to 4.536%).

[0103] Step S4: obtaining an actual mica crystal growth image; superimposing the temperature compensation area and the fluorine compensation area according to the current timestamp, and performing similarity matching with the actual mica crystal growth image, thereby displaying a mica material whole process visualization graph and generating a mica material whole process report;

[0104] Preferably, step S4 comprises:

[0105] Step S41: obtaining an actual mica crystal growth image;

[0106] Step S42: superimposing the temperature compensation area and the fluorine compensation area according to the current timestamp to obtain a mica crystal nucleus density prediction heat map;

[0107] Step S43: performing similarity matching between the mica crystal nucleus density prediction heat map and the actual mica crystal growth image, if the spatial similarity is greater than 0.9, then displaying a mica material whole process visualization graph and constructing a mica material whole process report.

[0108] In the embodiments of the present application, by installing a high-resolution industrial camera on the observation window of the kiln or the growth chamber, a digital image (usually a gray-scale or RGB pixel matrix, with a resolution of 1920x1080) of the current moment (with an accurate timestamp, for example, 2023-07-25T14:30:25.120Z) of the mica crystal growth process is captured in real time. This image is the "actual mica crystal growth image", which directly reflects the real state of the crystal morphology and crystal nucleus distribution at that moment.

[0109] Extract the "temperature compensation area" (for example, the kiln axial position interval [12.5m, 14.5m] needs to be cooled) and "fluorine compensation area" (for example, the entire batch fluorine content needs to be increased by +0.036%) calculated by the previous step (such as S3 or S5). The key operation is:

[0110] 1) Spatial mapping: convert the physical coordinates (kiln meters) of the temperature compensation area to pixel coordinates (such as [500px, 580px]) on the image according to the preset correspondence (for example, 1 meter = 40 pixels);

[0111] 2) Effect modeling: based on the fluorine compensation amount (ΔF), use the known influence model of fluorine content on crystal nucleus formation rate (such as fluorine increase promoting nucleation) to predict the possible change of crystal nucleus density in space (especially in the temperature compensation area) under this compensation;

[0112] 3) Heat map generation: combine the mapped temperature compensation area position and the fluorine compensation effect prediction to generate a "mica crystal nucleus density prediction heat map" in the image coordinate system.

[0113] The mica crystal nucleus density prediction heat map is essentially a two-dimensional matrix (same size as the image), and the value of each pixel point in the matrix represents the predicted crystal nucleus density level at that position (for example, in the range of 0-1, the higher the value, the more dense the predicted crystal nucleus), which is usually visualized using a color gradient (such as blue-yellow-red), and the red area represents the predicted high density.

[0114] Compare the generated prediction heat map (prediction matrix) with the actual growth image (real pixel matrix) obtained in step S41. The specific method is:

[0115] 1) Image preprocessing: possibly convert the actual image to a binary image or gradient image that highlights the crystal nucleus features;

[0116] 2) Similarity calculation: Calculate the similarity between the predicted heat map and the processed actual image in spatial (i.e. pixel position). For example, calculate the correlation coefficient of pixel values or the area ratio of the overlapping region between two images in the overlapping region (especially the temperature compensation region), to get a quantitative "spatial similarity" score (range 0-1). Judgment rule: if the calculated spatial similarity score is greater than the preset high threshold (0.9), it means that the predicted crystal nucleus distribution (based on the compensation scheme) is highly consistent with the actual observation, the compensation scheme is effective and the process state is consistent with the expectation.

[0117] Integrate the currently captured actual image, the superimposed displayed predicted heat map, the temperature / fluorine compensation region mark, the batch parameters (quality coefficient, fluorine content, etc.) and the historical characteristic data to generate a "mica material whole process visualization figure" containing key data and visual elements for the operator, and automatically output a structured "mica material whole process report", which contains a timestamp, compensation parameters, similarity value, key quality indicators, etc.

[0118] In one implementation manner of the embodiment of the present application, the actual crystal growth image (resolution 1920x1080) is captured at the timestamp 2023-07-25T14:30:25.120Z.

[0119] Temperature compensation region: physical position [10.0m, 12.0m] -> mapped to image horizontal axis pixel interval [400px, 480px] (assuming 1m = 40px). Fluorine compensation region: fluorine content increased by +0.036% (predicted to promote crystal nucleus formation). Predicted heat map: in the pixel interval [400, 480], according to the fluorine compensation effect model, it is predicted that the crystal nucleus density in this region will be 20% higher than the surrounding, and a heat map is generated (for example, displayed as a yellow to red transition in the [400, 480] interval).

[0120] After processing the actual image, the crystal nucleus position is identified and a density distribution map is generated. Calculate the overlap area ratio of the predicted heat map ([400, 480] region predicted high density) and the actual crystal nucleus density map in the [400, 480] region: 85% of the area of the predicted high density region is indeed observed in the actual image as a high density crystal nucleus, and the spatial similarity is calculated as 0.92 (greater than the threshold 0.9). Result: determine that the matching is successful. The system displays a visualization figure: the background is the actual growth image, and a red prediction box (temperature compensation region) is superimposed in the [400, 480] pixel interval, and a semi-transparent heat map color block is used to represent the predicted density distribution (basically coincides with the actual high crystal nucleus region). The whole process report records: timestamp 14:30:25.120Z, temperature compensation interval [10.0m, 12.0m] (pixels [400, 480]), fluorine compensation +0.036%, similarity 0.92, current batch Q value 0.83.

[0121] Preferably, step S43 comprises:

[0122] Step S431: Similarity matching is performed between the mica nucleus density prediction thermal map and the actual mica crystal growth image. If the spatial similarity degree is greater than 0.9, the thermal image pixel points are extracted to obtain the mica nucleus pixel points. The isosurface extraction is performed based on the isosurface points of the mica nucleus pixel points to obtain the mica thermal map vector data.

[0123] Step S432: The same isosurface is marked according to the mica thermal map vector data to obtain the mica material visualization graph.

[0124] Step S433: Affine change is performed on the mica material visualization graph, and the whole-process tracing of the production line is performed to obtain the mica material whole-process report.

[0125] In the embodiment of the application, when the spatial similarity of the prediction thermal map and the actual crystal growth image is determined to be greater than 0.9 (high matching), the following operations are performed: pixel point data is extracted from the prediction thermal map. The prediction thermal map is essentially a two-dimensional matrix (for example, 1920x1080) with the same size as the image, and the numerical value of each pixel point in the matrix represents the predicted crystal nucleus density level (for example, 0.0 to 1.0) at the position. All pixel points with a numerical value higher than a certain threshold (for example, 0.7) are screened out, and these points are considered as the predicted high crystal nucleus density area, referred to as “mica nucleus pixel points”. Then, the isosurface extraction is performed on these discrete “mica nucleus pixel points”: using an image processing algorithm (such as Marching Squares or its variants) to find out the continuous contour line (isoline) formed by all points with the same predicted density value. These contour lines are connected to outline the boundaries of the predicted high-density crystal nucleus region, the medium-density region and the low-density region. The boundary data (a polygon or curve composed of point coordinate sequences) is stored in a structured vector format (such as SVG path data or coordinate point list), which is “mica thermal map vector data”.

[0126] The vector data is used to mark the visualization graph: the extracted isosurface (i.e., the contour boundary of different predicted density regions) is superimposed and displayed on the original actual mica crystal growth image. Usually, different colors of lines or translucent color blocks are used to distinguish different density levels (for example, the area surrounded by red contour lines represents the predicted high-density area, and blue represents the low-density area), so that the “mica material visualization graph” which intuitively displays the comparison between the predicted crystal nucleus distribution area and the real growth image is generated.

[0127] For full-process traceability, the system will perform an affine transformation on this visualization: according to the known mapping relationship (such as scale, translation, rotation parameters) between the kiln coordinate system and the image pixel coordinate system, the pixel coordinates in the image (such as the contour point coordinates of a high-density area) are accurately converted back to the corresponding physical kiln position coordinates (in meters). Then, based on this physical position coordinate, the entire production process is traced back: the system automatically associates and extracts the relevant data generated by this position in all previous steps (S1, S2, S3…), such as the original temperature and pressure data of this position, the calculated lattice distortion value, the viscosity value, whether it is marked as a defect, compensation decision information, raw material batch information, etc.

[0128] Integrating all these traceability data and the current visualization, a structured "mica material full-process report" is generated. This report not only contains the final graph, but also lists in detail the key process parameters, calculated indicators, decision points, and results that affect the crystal growth in this area according to physical location or time sequence, providing complete production history traceability.

[0129] In one implementation of an embodiment of the present application, similarity = 0.92 > 0.9, triggering subsequent operations. In the predicted heat map, the pixel point (420, 300) has a value of 0.85, (430, 310) has a value of 0.82, and so on (all high-density points). Set the density threshold = 0.7, extract all pixel points with a value >= 0.7 (e.g. 5000 points in total). Isosurface extraction: the algorithm finds a closed contour line A (vector data: point sequence [(400, 280), (405, 285),..., (450, 310),..., (400, 280)]) that encloses all high-density points (value ≈ 0.8) with pixel coordinates roughly in the range (400-450, 280-320).

[0130] On the original actual growth image, draw contour line A with red lines to enclose the predicted high-nuclei-density area. Generate a visualization: the background is the actual crystal photo, superimposed with the red predicted high-density area contour.

[0131] Affine transformation: known mapping X(m) = 0.025 * u(pixel), Y(m) = 0.025 * v(pixel). Point (420, 300) on contour A is transformed to physical position (10.5 m, 7.5 m). Line trace back: system is positioned at 10.5 m from the furnace (belongs to zone 3): trace back S1 : temperature gradient at this position = 0.92 °C / cm, pressure spectrum integral = 0.38 kPa, lattice distortion = 0.22%. Trace back S2 / S3: melt viscosity at this position = 1150 Pa-s, not marked as defect. Trace back compensation decision: this zone belongs to temperature compensation zone (2.5 °C decrease needed), fluorine compensation +0.036%. Report generation: full process report contains: visual graphics (actual image with predicted contours). Key trace back data (position 10.5 m): S1 features: T_grad = 0.92, P_f = 0.38, ε = 0.22% S2 / S3 status: Viscosity = 1150 Pa-s, NoDefect S3 compensation: ΔΤ = -2.5 °C, ΔF = +0.036% Timestamp, batch ID, quality factor (Q = 0.83), similarity (0.92) and other global information.

[0132] Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the description given above, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

[0133] The foregoing is considered as illustrative only of the principles of the application. Numerous modifications and adaptations will be apparent to those skilled in the art in view of the above descriptions and the scope of the application is not to be limited to the examples shown but is to be given the full scope of the appended claims and equivalents thereof.

Claims

1. A digital visualization control method for a synthetic mica material production line, characterized in that, Includes the following steps: Step S1: Obtain industrial data of mica materials and extract data feature vectors to obtain batch quality coefficients, temperature-pressure fitting data and lattice distortion data; Step S2: Using the temperature-pressure fitting data along the kiln direction as the abscissa and the lattice distortion data as the axial coordinate, and combining the batch quality coefficient, output the mica material defect distribution map; Step S3: Determine the cold temperature zone based on the mica material defect distribution map. If the density difference of the cold temperature zone is greater than the preset density difference threshold, perform axial temperature compensation to obtain the temperature compensation area. Calculate the pinhole probability based on the mica material defect distribution map. If the pinhole defect probability is greater than the preset defect probability threshold, perform fluorine compensation to obtain the fluorine compensation area. Step S4: Obtain the actual mica crystal growth image; overlay the temperature compensation region and fluorine compensation region according to the current timestamp, and perform similarity matching with the actual mica crystal growth image to display a full-process visualization of mica materials and generate a full-process report of mica materials. Step S5: Generate control instructions for the mica material production line based on the mica material full-process report, and perform full-process control on the mica material production line according to the control instructions.

2. The digital visualization control method for a synthetic mica material production line as described in claim 1, characterized in that, Step S1 includes: Acquire industrial data for mica materials, including raw material feature vectors, thermal tensor data, and intrinsic crystal growth properties; the thermal tensor data includes the axial temperature gradients of thermocouples in six temperature zones and the pressure pulsation spectrum of pressure transmitters. The batch quality coefficient is obtained by encoding the feature vector of raw materials with fluorine content as a benchmark. Wavelet fusion is performed on the axial temperature gradient and pressure pulsation spectrum of the thermal tensor data to output temperature-pressure fitting data. By analyzing the biaxial interference of the intrinsic eigenvalues ​​of crystal growth, lattice distortion data can be obtained.

3. The digital visualization control method for a synthetic mica material production line as described in claim 2, characterized in that, The biaxial interference of wavelet fusion and analysis of crystal growth eigenvalues ​​based on the axial temperature gradient and pressure pulsation spectrum of thermal tensor data includes: Six temperature zone acquisition points were set along the longitudinal axis of the kiln, and envelope extraction was performed on the six temperature zone acquisition points to obtain the axial temperature gradient; the pressure spectrum energy of the kiln was integrated by signal amplitude to obtain the pressure pulsation spectrum; The axial temperature gradient and pressure pulsation spectrum are fused using wavelet fusing and sorted according to temperature-pressure to obtain temperature-pressure fitting data; The thickness of the mica crystal was obtained; the extraordinary and ordinary refractive indices were extracted from the intrinsic properties of crystal growth, and the birefringence compensation data were calculated using the biaxial interference analytical formula, which is as follows: Δδ=(n e -n o )d-d; Where Δδ is the birefringence compensation data, n e For the refractive index of light, n o denoted as the ordinary refractive index, d as the mica crystal thickness, and δ as the mica birefringence. The optical path is simulated using birefringence compensation data, and the lattice distortion data is obtained by compensating for the difference based on the simulated optical path.

4. The digital visualization control method for a synthetic mica material production line as described in claim 1, characterized in that, Step S2 includes the following steps: Step S21: Construct a spatial mica evolution coordinate system by unfolding the temperature-pressure fitting data along the kiln direction as the abscissa and using the lattice distortion data as the axial coordinate; Step S22: Locate the crystal phase region of the spatial mica evolution coordinate system, identify the suspended mica crystal nuclei in the crystal phase region, and generate mica crystal nuclei identification data; mark the mica crystal nuclei identification data according to the fluorine content, and thus output the mica melt viscosity distribution map; Step S23: Based on the batch quality coefficient, find the production line defects in the mica melt viscosity distribution map to obtain the mica material defect distribution map.

5. The digital visualization control method for a synthetic mica material production line as described in claim 4, characterized in that, Step S23 includes: Based on the viscosity distribution map of mica melt, the reaction rate and viscosity standard deviation were extracted to obtain mica reaction rate data and viscosity standard deviation data. Obtain a historical defect database; perform cross-correlation processing on the mica melt viscosity distribution map based on the mica fluorine content of the batch quality coefficient; when the mica reaction rate data is less than 0.8 and the viscosity standard deviation data is greater than 50 Pa·s, use the historical defect database to mark production defects and obtain mica production line marking data. Based on the statistical defect probability of mica production line marking data, when the defect rate of a single production line in the mica production line marking data is greater than 0.2, a defect distribution map of mica material is constructed.

6. The digital visualization control method for a synthetic mica material production line as described in claim 1, characterized in that, Step S3 includes: The deviation of the longitudinal mica nucleus density is calculated based on the mica material defect distribution map. If the nucleus density deviation is greater than 15% of the preset density difference threshold, temperature compensation is performed using the axial temperature compensation formula to obtain the temperature compensation area. The quality coefficient of the current batch is extracted based on the defect distribution map of mica material. If the probability of pinhole defects in the quality coefficient of the current batch is greater than the preset defect probability threshold of 30%, a new formula is set based on the new fluorine content formula setting formula to obtain the fluorine compensation area.

7. The digital visualization control method for a synthetic mica material production line as described in claim 6, characterized in that, The axial temperature compensation formula and the formula for setting the new fluorine content formula include: Where ΔT is the axial temperature compensation, t1 and t2 are the start and end times of the integration time interval, f(Δρ) is the density deviation function, and Δρ is the density difference of mica crystal nuclei. ΔF=0.2×|K_batch 1 -1|; Where ΔF is the data set for the new fluorine content formula, and K_batch 1 This represents the quality coefficient for the current batch.

8. The digital visualization control method for a synthetic mica material production line as described in claim 1, characterized in that, Step S4 includes: Step S41: Obtain actual mica crystal growth images; Step S42: Overlay the temperature compensation region and the fluorine compensation region according to the current timestamp to obtain the predicted thermal map of mica crystal nucleus density; Step S43: Perform similarity matching between the predicted thermal map of mica crystal nucleus density and the actual mica crystal growth image. If the spatial similarity is greater than 0.9, display the full-process visualization graphics of mica materials and construct a full-process report of mica materials.

9. The digital visualization control method for a synthetic mica material production line as described in claim 8, characterized in that, Step S43 includes: Step S431: Perform similarity matching between the predicted mica nucleus density heat map and the actual mica crystal growth image. If the spatial similarity is greater than 0.9, extract the pixels from the heat map to obtain the mica nucleus pixels. Extract the isosurface based on the isovalues ​​of the mica nucleus pixels to obtain the mica heat map vector data. Step S432: Mark the same isosurfaces based on the mica thermal map vector data to obtain a visualization graphic of the mica material; Step S433: Perform affine transformations on the visualized graphics of mica materials and trace the entire production line process to obtain a full-process report on mica materials.

10. A digital visualization control system for a synthetic mica material production line, characterized in that, The method for implementing the digital visualization control method for a synthetic mica material production line as described in claim 1 includes: The data acquisition and feature construction module is used to acquire industrial data of mica materials and extract data feature vectors to obtain batch quality coefficients, temperature-pressure fitting data and lattice distortion data. The defect modeling and probability analysis module is used to output a defect distribution map of mica material by unfolding temperature-pressure fitting data along the kiln direction as the horizontal axis and lattice distortion data as the axial axis, and combining the batch quality coefficient. The defect compensation strategy generation module is used to determine the cold temperature zone based on the defect distribution map of mica material. If the density difference of the cold temperature zone is greater than the preset density difference threshold, axial temperature compensation is performed to obtain the temperature compensation area. The module also calculates the pinhole probability based on the defect distribution map of mica material. If the pinhole defect probability is greater than the preset defect probability threshold, fluorine compensation is performed to obtain the fluorine compensation area. The visualization and report generation module is used to acquire actual mica crystal growth images; temperature compensation regions and fluorine compensation regions are superimposed according to the current timestamp, and similarity matching is performed with the actual mica crystal growth images to display a full-process visualization of mica materials and generate a full-process report of mica materials. The control instruction module is used to generate control instructions for the mica material production line based on the full-process report of mica materials, and to perform full-process control of the mica material production line according to the control instructions.

Images