Intelligent melting control system of quartz crucible melting machine

Abstract

The invention relates to the technical field of quartz crucible manufacturing, and particularly discloses a quartz crucible melting machine melting intelligent control system, which comprises a precision correction module, a preheating heating module, a melting soaking module and a clarification constant temperature module. The precision correction module carries out cross validation through a double-color thermodetector and predicted temperature based on arc length and power, and online calibration of temperature measurement equipment is achieved. The preheating temperature rising module monitors the surface temperature rising rate of quartz sand and performs dynamic regulation and control; the melting soaking module adopts a polar coordinate grid method to distribute points, judges the heating uniformity of the melt and identifies local hot spots, local cold spots or an overall mode problem, so that a clear basis is provided for intervention; and the clarification constant temperature module monitors a temperature curve in real time, identifies a fluctuation mode and traces reasons. According to the invention, full-flow temperature intelligent control is realized, the temperature measurement precision, the melt uniformity and the finished product quality are improved, and the method is suitable for a quartz crucible melting production environment.

Description

Technical Field

[0001] This invention relates to the field of quartz crucible manufacturing technology, and specifically to an intelligent control system for a quartz crucible melting machine. Background Technology

[0002] Quartz crucibles are a key basic material in semiconductor, photovoltaic and other fields, and their quality directly affects the performance and yield of downstream products. The melting process of quartz crucibles is the most critical and precise link in the entire production process, especially the temperature control during the melting stage, which has a decisive impact on the purity, uniformity and structural integrity of the final product.

[0003] Currently, there are still significant shortcomings in temperature control during the quartz crucible melting process: First, the temperature measurement methods are limited, relying mainly on single-point or single-band temperature measurement equipment, which is easily affected by factors such as contamination of the observation window, target obstruction, and changes in distance, resulting in large temperature measurement errors. Furthermore, there is a lack of effective cross-validation mechanisms, making it difficult to guarantee the reliability of the data.

[0004] Secondly, existing control systems often focus on temperature regulation at a certain stage, lacking the ability to dynamically track and intelligently control the entire process from preheating and melting to clarification. This makes it difficult to accurately judge and intervene in the uniformity of melting heating and temperature stability in real time, which can easily lead to defects such as local thermal stress cracks, uneven melt, and residual bubbles, affecting the quality of finished products and production consistency.

[0005] Therefore, there is an urgent need to develop an intelligent control system for quartz crucible melting machines that can achieve cross-validation of multi-source temperature measurement data, dynamic temperature control throughout the entire process, and intelligent diagnosis and intervention capabilities, in order to improve the stability of the melting process and the quality of the finished product. Summary of the Invention

[0006] To address the above problems, this invention proposes an intelligent control system for a quartz crucible melting machine, which enables the regulation of the melting temperature of the quartz crucible.

[0007] The technical solution adopted by the present invention to solve its technical problem is as follows: The present invention provides a quartz crucible melting machine intelligent control system, including: a precision correction module, used to verify the reliability of the temperature measurement data and calibrate the dual-color thermometer by comparing the heating temperature detected by the dual-color thermometer during the trial heating process with the heating temperature predicted based on the arc length and power before melting.

[0008] The preheating module is used to monitor the heating rate of the quartz sand surface in real time during the preheating stage, and to make corresponding adjustments based on the difference between the heating rate and the expected heating rate.

[0009] The melting and homogenizing module is used to select multiple temperature measurement points on the surface of the melt using the polar coordinate grid method during the melting stage. It judges the heating uniformity of the melt based on the temperature difference between the measurement points, and identifies and intervenes in case of uneven heating. The problem types include local hot spots, local cold spots, and overall pattern problems.

[0010] The clarification and isothermal module is used to monitor the melt temperature in real time during the clarification stage and generate a temperature curve. Based on the temperature curve, it determines whether the temperature fluctuation exceeds the allowable range, and if it exceeds the limit, it traces the suspected cause based on the temperature fluctuation pattern. The temperature fluctuation pattern includes regular oscillation, irregular jump, unidirectional drift and step-like jump.

[0011] Compared with the prior art, the intelligent control system for quartz crucible melting machine described in this invention has the following advantages: 1. This invention achieves online calibration of the temperature measuring equipment by cross-validating the predicted temperature based on arc length and power using a dual-color thermometer, effectively reducing environmental interference and improving the accuracy and consistency of temperature data.

[0012] 2. This invention covers the three stages of the melting process: preheating, melting, and clarification. It achieves intelligent temperature control throughout the entire process from initial heating to final molding by functions such as monitoring the heating rate, judging the uniformity of the polar coordinate grid, and recognizing the temperature fluctuation pattern.

[0013] 3. This invention uses a polar coordinate grid layout to simultaneously analyze circumferential temperature difference and radial temperature gradient, comprehensively capture abnormal temperature distribution, enhance the ability to judge the uniformity of melt heating, accurately identify local hot spots, local cold spots or overall pattern problems, and provide a clear basis for intervention. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a system module connection diagram of the present invention.

[0016] Figure 2 This is a schematic diagram of the principle of melting quartz crucibles using the electric arc centrifugal method of the present invention.

[0017] Figure 3 This is a schematic diagram of the layout of temperature measuring points on the melt surface according to the present invention.

[0018] Reference numerals in the attached figures: 1. Observation window; 2. Temperature measuring port; 3. Quartz crucible; 4. Graphite mold; 5. Heating component; 6. Electrode; 7. Crucible lifting and rotating mechanism. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, the present invention provides an intelligent control system for a quartz crucible melting machine, including a precision correction module, a preheating module, a melting homogenization module, and a clarification and constant temperature module.

[0021] The preheating module is connected to the accuracy correction module and the melting and homogenizing module, respectively, and the clarification and constant temperature module is connected to the melting and homogenizing module.

[0022] The accuracy correction module is used to verify the reliability of the temperature measurement data and calibrate the dual-color thermometer by comparing the heating temperature detected by the dual-color thermometer during the trial heating process with the heating temperature predicted based on the arc length and power before melting.

[0023] Furthermore, the specific working process of the accuracy correction module is as follows: multiple temperature measuring ports are set around the furnace body and dual-color thermometers are arranged to form a temperature measuring network.

[0024] Select a test temperature point at the junction of the inner wall and bottom of the graphite mold.

[0025] Before melting, a trial heating was performed to obtain the heating temperature of each test point detected by a dual-color thermometer during the trial heating process.

[0026] The arc image during the trial heating is acquired through the observation window. The arc length is obtained through image processing technology. The arc length is then input into the arc length-heating temperature correlation model to obtain the heating temperature predicted based on the arc length.

[0027] The arc power during the trial heating period is retrieved from the console and input into the arc power-heating temperature correlation model to obtain the heating temperature predicted based on the arc power.

[0028] The heating temperatures predicted based on arc length and arc power are weighted and fused to obtain a comprehensive predicted heating temperature.

[0029] Based on the preset relationship between graphite mold thickness and temperature decay, the temperature decay is determined in combination with the current graphite mold thickness.

[0030] The predicted heating temperature of the test point is calculated by subtracting the temperature decay from the comprehensive predicted heating temperature.

[0031] The heating temperature of each test point of the dual-color thermometer is compared with the predicted heating temperature of the test point to obtain the temperature measurement deviation of each dual-color thermometer.

[0032] If the temperature measurement deviation is less than the set threshold, the temperature measurement data of the corresponding dual-color thermometer is determined to be reliable; otherwise, it is determined to be unreliable.

[0033] Calibrate the two-color thermometers that are deemed unreliable.

[0034] It should be noted that, for reference Figure 2 As shown, the present invention uses an electric arc centrifugal method to melt the quartz crucible.

[0035] It should be noted that the present invention uses a dual-color thermometer to measure temperature. The dual-color thermometer calculates temperature by measuring the ratio of the radiation energy of two similar wavelengths, which can effectively reduce interference from factors such as contamination of the observation window, partial obstruction of the target, and changes in measurement distance. It is more suitable for the complex industrial environment of quartz crucible melting than a single-band thermal imager.

[0036] It should be noted that if the electric arc fluctuates unstablely during the trial heating process, the average arc length should be taken as the final arc length.

[0037] It should be noted that the specific method for obtaining the correspondence between the thickness of the graphite mold and the temperature decay is as follows: by conducting a standard heating experiment on graphite molds of different thicknesses, simultaneously measuring the temperature on the heat source side and the temperature at the test point inside the mold, calculating the difference between the two to obtain the actual temperature decay, and then fitting the correspondence between the thickness of the graphite mold and the temperature decay, wherein the correspondence is a curve or a mathematical model.

[0038] It should be noted that the weights of the weighted fusion can be preset based on experience in the field, or obtained through a limited number of experimental data. For example, historical data of actual heating temperature under different combinations of arc length and arc power can be collected first, and then the correlation coefficients of arc length, arc power and heating temperature can be calculated separately. Regression analysis can be used to determine the contribution of the two to the prediction of heating temperature. Finally, after normalization, the contribution is converted into a weight coefficient and the sum of the two weights is 1.

[0039] For example, in one specific embodiment, the following steps can be performed to determine weights in a data-driven manner: collect no fewer than 50 sets of historical test heating data, each set of data including arc length, arc power and calibrated reference temperature.

[0040] The correlation coefficients between arc length, arc power, and reference temperature were calculated separately. Assuming a correlation coefficient of 0.85 for length and 0.70 for power, multiple linear regression analysis yielded standardized regression coefficients of 0.6 and 0.4 for length and power, respectively. After normalization, the weights were preset as follows: arc length weight 0.6, arc power weight 0.4. These weights can be recalibrated every 100 crucibles produced, based on electrode wear or process changes.

[0041] It should be noted that the temperature measurement deviation threshold is set by analyzing historical calibration data or conducting special tests. Specifically, this involves collecting a large amount of temperature measurement deviation data of the dual-color thermometer during trial heating, combining it with the maximum temperature measurement error allowed by the process requirements, and using statistical methods such as calculating confidence intervals to determine the upper limit of the threshold, ensuring that the threshold can effectively distinguish between normal measurement errors and abnormal deviations.

[0042] Specifically, the threshold is a dynamic value based on statistical analysis and process requirements. For example, temperature deviation data from over 200 trial heating processes can be collected, and their mean and standard deviation can be calculated. Assuming the deviation data follows a normal distribution with a mean of 0°C and a standard deviation of 5°C, and the maximum allowable temperature measurement error is ±15°C according to process requirements, ±15°C can be used as a fixed threshold. More preferably, the 3σ principle can be adopted, setting the threshold to the mean ± 3 times the standard deviation, i.e., ±15°C. As equipment stability improves, this threshold can be tightened to ±2σ (i.e., ±10°C) for higher accuracy.

[0043] It should be noted that the basis for cross-validating the heating temperature predicted by arc length and power with the values ​​detected by the dual-color thermometer is that there is a clear mechanistic relationship between the physical parameters of the arc and the heating temperature, which can provide a temperature estimation benchmark independent of optical measurements. Its function is to identify whether the dual-color thermometer has deviations or malfunctions through comparison of multi-source data, thereby achieving online verification and accurate calibration of the temperature measuring instrument, ensuring the reliability and consistency of temperature monitoring data during the melting process.

[0044] In this embodiment, the present invention cross-validates the temperature measurement equipment by using a dual-color thermometer and a predicted temperature based on arc length and power, thereby achieving online calibration of the temperature measurement equipment, effectively reducing environmental interference, and improving the accuracy and consistency of temperature data.

[0045] Furthermore, the process of establishing the arc length-heating temperature correlation model is as follows: retrieve arc images collected during multiple historical melting cycles and synchronously recorded melt temperature data from the database.

[0046] The arc image is processed to extract the arc length.

[0047] The extracted arc length is used as the input variable, and the melt temperature recorded at the corresponding time is used as the output variable.

[0048] Regression analysis was used to fit the input and output variables, establish the mapping relationship between arc length and heating temperature, and construct an arc length-heating temperature correlation model.

[0049] It should be noted that the melt temperature was obtained by measuring a calibrated two-color thermometer.

[0050] Furthermore, the process of establishing the arc power-heating temperature correlation model is as follows: retrieve arc current and voltage data collected in multiple historical melting cycles and melt temperature data recorded synchronously from the database.

[0051] The arc power is calculated based on the current and voltage, and the obtained power data is preprocessed.

[0052] The pre-processed arc power is used as the input variable, and the melt temperature recorded at the corresponding time is used as the output variable.

[0053] Regression analysis was used to fit the input and output variables, establish the mapping relationship between arc power and heating temperature, and construct an arc power-heating temperature correlation model.

[0054] It should be noted that the preprocessing of the electrical power data includes filtering and normalization to eliminate noise interference and unify the data dimensions.

[0055] The preheating module is used to monitor the heating rate of the quartz sand surface temperature in real time during the preheating stage, and to make corresponding adjustments based on the difference between the heating rate and the expected heating rate.

[0056] Furthermore, the specific working process of the preheating module is as follows: during the preheating stage, several detection points are selected on the edge line formed by the contact between the horizontal surface of the quartz sand and the inner wall of the graphite mold.

[0057] The temperature at each detection point is collected in real time, and the average value is calculated as the surface temperature of the quartz sand.

[0058] The heating rate of the surface temperature is monitored and compared with the set expected heating rate to obtain the heating rate deviation.

[0059] If the current heating rate deviation exceeds the preset allowable range, and the deviation at the next moment still exceeds the range, then it is determined that the heating rate needs to be adjusted.

[0060] Based on the heating rate deviation, the adjustment direction and amount of the heating rate are determined and adaptive control is performed.

[0061] It should be noted that the expected heating rate is set based on the material properties of the quartz sand, the target melting process requirements, and historical production data. Specifically, it is determined by analyzing the temperature change rate in the preheating stage of the ideal process curve and combining it with expert experience values ​​to ensure that the quartz sand is heated uniformly and avoids thermal stress cracking.

[0062] For example, for quartz sand with a purity of 99.99% or higher and a particle size distribution of 180-250 μm, the ideal preheating curve requires a smooth transition to the melting point. Analysis of historical data from 100 successful production batches reveals that the heating rate during the preheating stage (e.g., from room temperature to 1200°C) is mainly concentrated in the range of 4.5°C / min to 5.5°C / min. Based on materials expert experience, to prevent thermal stress cracking, the heating rate should not exceed 6°C / min. Therefore, the expected heating rate can be set at 5.0°C / min. If a coarser or finer quartz sand raw material is used, this value needs to be fine-tuned to 4.8°C / min or 5.2°C / min accordingly.

[0063] It should be noted that the allowable range of heating rate deviation is obtained by statistically analyzing the fluctuation range of heating rate during the preheating stage in normal production processes. The confidence interval calculation method is adopted, and reasonable upper and lower limits of deviation are set by comprehensively considering process tolerance requirements and equipment control accuracy.

[0064] The specific setting method is as follows: Select 100 normal production batches and extract the heating rate data per minute during the preheating stage, then calculate the standard deviation (σ) of the overall data. Assume the calculated σ = 0.3°C / min. Considering the inherent slight fluctuations in the system, the allowable range can be set to the expected value ±2σ, i.e., ±0.6°C / min. This means that when the real-time heating rate continuously exceeds the range of 4.4°C / min to 5.6°C / min, the system will trigger regulation. This range can be fine-tuned according to the temperature control performance of different furnaces; for high-performance furnaces, it can be tightened to ±0.4°C / min.

[0065] It should be noted that the heating rate during the preheating stage is controlled to ensure that the quartz sand layer heats up uniformly and steadily, preventing local thermal stress cracks caused by excessively rapid heating, or low energy efficiency and uneven heating of materials caused by excessively slow heating. This provides a stable and controllable preheating foundation for the subsequent melting stage, avoids adverse effects of sudden temperature changes on melt quality, and thus ensures the finished product quality and production consistency of the quartz crucible.

[0066] In this embodiment, the present invention avoids energy waste and material thermal stress damage by adaptively controlling the heating rate during the preheating stage, ensuring a stable melting process and ultimately improving the finished product quality of the quartz crucible and the consistency between production batches.

[0067] Furthermore, the specific process for determining the adjustment direction and amount of the heating rate is as follows: the adjustment direction is determined based on the sign of the heating rate deviation.

[0068] If the deviation is positive, it indicates that the current heating rate is greater than the expected heating rate, and the direction of adjustment for the heating rate is to decrease it.

[0069] If the deviation is negative, it indicates that the current heating rate is less than the expected heating rate, and the direction of adjustment for the heating rate is to increase it.

[0070] The absolute value of the heating rate deviation is used as the adjustment amount of the heating rate.

[0071] The melting and homogenizing module is used to select multiple temperature measurement points on the surface of the melt using the polar coordinate grid method during the melting stage. It judges the heating uniformity of the melt based on the temperature difference between the measurement points, and identifies the problem type and intervenes when the heating is uneven. The problem types include local hot spots, local cold spots and overall pattern problems.

[0072] Furthermore, the specific working process for determining the heating uniformity of the melt in the melting homogenization module is as follows: (See...) Figure 3 As shown, concentric rings are set with a fixed radius step size, with the center point of the melt surface as the origin, and rays are drawn from the origin with a fixed angular step size. Temperature measuring points are set at the intersection of the concentric rings and the rays.

[0073] The temperature of the center point of the melt surface and each measuring point is collected in real time.

[0074] Calculate the maximum temperature difference between all temperature measuring points on the same ring, and use it as the relative temperature difference on the ring.

[0075] Calculate the temperature change between temperature measurement points at different radii along the same ray, and construct a radial temperature gradient sequence for that ray.

[0076] If the following conditions are met simultaneously, the melt is determined to be heated uniformly: (1) The relative temperature difference on all the rings is less than the set relative temperature difference threshold on the rings.

[0077] (2) The matching degree between the radial temperature gradient sequence of all rays and the reference radial temperature gradient sequence is greater than the set matching degree threshold.

[0078] Otherwise, it is judged as uneven heating.

[0079] It should be noted that the relative temperature difference threshold on the ring is determined based on the material thermal uniformity process requirements and historical normal production data statistics. By analyzing the maximum temperature difference distribution on each ring under a large number of uniformly heated conditions, the upper limit of its confidence interval is taken as the threshold to ensure that the uniformity and abnormality of the circumferential temperature distribution can be effectively distinguished.

[0080] For example, to set this threshold, a large amount of production data (e.g., 50 heats) under uniform heating conditions can be collected, and the maximum temperature difference between all temperature measurement points on each ring can be calculated. Assume that for a specific ring of radius R, the 95th percentile of these maximum temperature differences is 8°C. To ensure process robustness, an additional margin of approximately 25% can be added, setting the relative temperature difference threshold on the ring to 10°C. The threshold can be different for rings of different radii; for example, the threshold could be set to 8°C for the inner ring (smaller radius), 10°C for the middle ring, and 12°C for the outer ring (larger radius) to match their inherent thermal field distribution characteristics.

[0081] It should be noted that the reference radial temperature gradient sequence is generated by collecting temperature gradient data of multiple ray directions under uniform heating conditions during historical melting processes, and then smoothing and averaging the data. It represents the reasonable variation law of radial temperature under ideal heating conditions.

[0082] A specific generation method is illustrated below: From 30 high-quality production batches with uniform heating, select 5 key points (e.g., radius ratios of 0, 0.25, 0.5, 0.75, 1.0) from the center to the edge on each ray. Calculate the average temperature difference of each point relative to the center point across all batches to obtain an initial gradient sequence, such as [0, -3, -7, -12, -15] (°C). Then, use a moving average method for smoothing to finally obtain a smooth reference radial temperature gradient sequence, such as [0, -2.5, -5.8, -10.5, -14] (°C). This sequence represents the ideal radial temperature distribution under this specific equipment and process.

[0083] It should be noted that the matching metric is quantified by calculating the similarity index between the real-time radial temperature gradient sequence and the reference sequence. The specific value of the matching degree is the Pearson correlation coefficient or cosine similarity. The higher the similarity, the greater the matching degree.

[0084] Taking the Pearson correlation coefficient as an example, its value ranges from [-1, 1]. Under uniform heating conditions, the correlation coefficient between the real-time gradient and the reference gradient is usually extremely high. Statistical analysis of historical data shows that this coefficient is generally greater than 0.92 under normal conditions. Therefore, the matching threshold can be set to 0.90. When the calculated correlation coefficient is lower than 0.90, it is determined that the radial temperature distribution in that ray direction is abnormal. If cosine similarity is used, its value range is [0, 1], and it is usually greater than 0.98 under normal conditions, so the threshold can be set accordingly to 0.975.

[0085] It should be noted that by synchronously analyzing the circumferential temperature difference and radial gradient through polar coordinate grids, anomalies in the surface temperature distribution of the melt in the angular and radial directions can be fully captured, improving the accuracy and reliability of uniformity judgment, providing a clear basis for subsequent precise intervention, and thus improving the consistency of melt quality.

[0086] In this embodiment, the present invention adopts a polar coordinate grid layout method to simultaneously analyze the circumferential temperature difference and radial temperature gradient, comprehensively capture abnormal temperature distribution, enhance the ability to judge the uniformity of melt heating, accurately identify local hot spots, local cold spots or overall pattern problems, and provide a clear basis for intervention.

[0087] Furthermore, the specific process for identifying the problem type when the heating is uneven in the melting and homogenizing module is as follows: calculate the average temperature of each temperature measuring point on the same ring to obtain the average temperature of the ring.

[0088] The average temperature of each ring is compared with the temperature at the center point of the melt surface to obtain the temperature difference.

[0089] If the temperature difference of all the rings exceeds their respective preset allowable range, it is determined to be an overall mode problem; otherwise, perform the following steps: number each temperature measurement point on the same ring in the set order, establish a coordinate system with the number as the abscissa and temperature as the ordinate, and draw a scatter plot of the temperature distribution of each ring.

[0090] The fluctuation bands of data points in the scatter plot are obtained by regression and fitting methods. If a data point deviates from the fluctuation band and its temperature is higher than the upper limit of the fluctuation band, the point is recorded as a local hot spot. If its temperature is lower than the lower limit of the fluctuation band, it is recorded as a local cold spot.

[0091] Count the number, location, and offset of local hot spots and local cold spots on all the rings.

[0092] It should be noted that the allowable range of temperature difference between each ring is obtained by statistically analyzing the range of difference between the average temperature of each ring and the temperature at the center point during the melting stage in the historical normal production process. The confidence interval calculation method is adopted, and the material thermal uniformity process requirements and system measurement errors are comprehensively considered to set reasonable upper and lower limits for the allowable range of each ring.

[0093] The specific setting process is as follows: For each ring, the temperature difference between it and the center point under normal production conditions is statistically analyzed. For example, for the first ring (the innermost ring), the temperature difference data is distributed between -4°C and +4°C; for the second ring, it is distributed between -7°C and +6°C; and for the third ring, it is distributed between -11°C and +9°C. Taking the 95% confidence interval and rounding appropriately, the allowable range for each ring can be set. For example, the allowable range for the first ring is ±5°C, for the second ring it is -6°C to +7°C (or a symmetrical value of ±7°C), and for the third ring it is -10°C to +10°C. This differentiated setting better reflects the actual thermal field distribution.

[0094] It should be noted that for identified local hot spots, measures are taken to reduce the energy input in the corresponding areas; for local cold spots, the energy supply in the corresponding areas is increased; if it is a problem of the overall mode, the heating system is calibrated globally and the process parameters are optimized.

[0095] The clarification and isothermal module is used to monitor the melt temperature in real time during the clarification stage and generate a temperature curve. Based on the temperature curve, it is determined whether the temperature fluctuation exceeds the allowable range. If the limit is exceeded, the suspected cause is traced according to the temperature fluctuation pattern. The temperature fluctuation pattern includes regular oscillation, irregular jump, unidirectional drift and step-like jump.

[0096] Furthermore, the specific working process of the clarification and isothermal module is as follows: during the clarification stage, the melt temperature is monitored in real time, and a temperature curve showing its change over time is plotted.

[0097] The linear regression line of the temperature curve is used as a baseline.

[0098] Identify each extreme point on the temperature curve, calculate the fluctuation of each extreme point relative to the baseline, mark the extreme points with fluctuations greater than a set threshold as fluctuation points, and record the time span corresponding to each fluctuation point as the fluctuation duration.

[0099] If the total number of fluctuation points and the cumulative duration of fluctuations do not exceed their respective set allowable ranges, the temperature fluctuation is determined to be within limits; otherwise, it is determined to be within limits, and the following steps are performed: determine the specific pattern of temperature fluctuation based on the morphological characteristics of the temperature curve, combine the pre-established association set of each fluctuation pattern with potential causes, match the suspected cause of the current temperature fluctuation, and provide feedback.

[0100] It should be noted that the allowable range of the total number of fluctuation points and the cumulative duration of fluctuations is obtained by statistically analyzing the fluctuation characteristics of the temperature curve during the historical normal clarification process. The distribution range of the number of fluctuation points and the duration of fluctuations are calculated respectively, and a reasonable threshold is set based on the confidence interval method combined with the process stability requirements.

[0101] For example, consider the following setting: Analyze the temperature profiles of the clarification stage (60 minutes) of 50 normal batches. Count the total number of fluctuation points (fluctuations exceeding ±3°C) for each curve; the 95th percentile is 3. Count the percentage of total fluctuation duration out of the total duration; the 95th percentile is 4.5%. Based on this, the allowable range can be set as follows: the total number of fluctuation points should not exceed 3, and the cumulative fluctuation duration should not exceed 5% of the total duration. For example, for a 60-minute clarification stage, the cumulative fluctuation duration should not exceed 3 minutes. This threshold can be adjusted according to product grade requirements; for high-specification products, it can be tightened to 2 points and 3% of the total duration.

[0102] It should be noted that the association set is established based on the analysis of a large amount of historical abnormal data and the summarization of process knowledge. By extracting the temperature curve features under different fluctuation modes and combining multi-source information such as equipment status, process parameters and external interference, causal association mining is carried out to form a mapping relationship library between each mode and potential causes.

[0103] In this embodiment, the present invention identifies fluctuation patterns such as regular oscillations and irregular jumps through temperature curve analysis during the clarification stage, and combines historical data and process knowledge to trace the root cause, thereby achieving rapid diagnosis and feedback and avoiding quality abnormalities.

[0104] Furthermore, the process of determining the specific temperature fluctuation pattern is as follows: extracting example temperature curves corresponding to each temperature fluctuation pattern pre-stored in the database, wherein the temperature fluctuation pattern includes regular oscillation, irregular jumps, unidirectional drift, and step-like jumps.

[0105] The temperature curves obtained from real-time monitoring during the clarification phase are compared with the sample temperature curves in terms of shape similarity.

[0106] The temperature fluctuation pattern corresponding to the example curve with the highest similarity is determined as the final identified fluctuation pattern.

[0107] It should be noted that each temperature fluctuation pattern typically exhibits distinct visual characteristics: regular oscillations are characterized by the temperature curve fluctuating around the set value in an approximately sinusoidal periodic manner, with its amplitude and period often being relatively stable; irregular jumps are manifested as sudden sharp rises or falls in temperature data, forming brief peaks or troughs, followed by rapid recovery, with drastic changes and no repeating patterns; unidirectional drift refers to the temperature measurement value continuously and slowly deviating from the set value, showing a monotonically rising or falling trend; step-like jumps refer to the temperature suddenly dropping or rising, and then reaching a steady state again at the new level, forming a clear step response.

[0108] In this embodiment, the present invention covers the three stages of the melting process: preheating, melting, and clarification. Through functions such as heating rate monitoring, polar coordinate grid uniformity judgment, and temperature fluctuation pattern recognition, it realizes intelligent temperature control throughout the entire process from initial heating to final molding.

[0109] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0110] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0111] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0112] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0113] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0114] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart control system for melting a quartz crucible melting machine, characterized in that, include: The accuracy calibration module is used to verify the reliability of the temperature measurement data and calibrate the dual-color thermometer by comparing the heating temperature detected by the dual-color thermometer during the trial heating process with the heating temperature predicted based on the arc length and power before melting. The preheating module is used to monitor the heating rate of the quartz sand surface in real time during the preheating stage, and to make corresponding adjustments based on the difference between the heating rate and the expected heating rate. The melting and homogenizing module is used to select multiple temperature measurement points on the surface of the melt using the polar coordinate grid method during the melting stage. It judges the uniformity of the melt heating based on the temperature difference between the measurement points, and identifies and intervenes when the heating is uneven. The problem types include local hot spots, local cold spots, and overall pattern problems. The clarification and isothermal module is used to monitor the melt temperature in real time during the clarification stage and generate a temperature curve. Based on the temperature curve, it determines whether the temperature fluctuation exceeds the allowable range. If the limit is exceeded, it traces the suspected cause based on the temperature fluctuation pattern. The temperature fluctuation patterns include regular oscillation, irregular jumps, unidirectional drift, and step-like jumps.

2. The intelligent control system for a quartz crucible melting machine according to claim 1, characterized in that: The specific working process of the accuracy correction module is as follows: Multiple temperature measuring ports are set around the furnace body and dual-color thermometers are arranged to form a temperature measuring network; Select a test temperature point at the junction of the inner wall and bottom of the graphite mold; Before melting, a trial heating was performed to obtain the heating temperature of each test point detected by a dual-color thermometer during the trial heating process; Arc images are captured during the trial heating process through an observation window. The arc length is obtained through image processing technology. This arc length is then input into the arc length-heating temperature correlation model to obtain the heating temperature predicted based on the arc length. The arc power during the trial heating period is retrieved from the console and input into the arc power-heating temperature correlation model to obtain the heating temperature predicted based on the arc power. The heating temperatures predicted based on arc length and arc power are weighted and fused to obtain a comprehensive predicted heating temperature; Based on the preset relationship between graphite mold thickness and temperature decay, the temperature decay is determined in combination with the current graphite mold thickness. The predicted heating temperature of the test point is calculated by subtracting the temperature decay from the comprehensive predicted heating temperature. The heating temperature of each test point of the dual-color thermometer is compared with the predicted heating temperature of the test point to obtain the temperature measurement deviation of each dual-color thermometer. If the temperature measurement deviation is less than the set threshold, the temperature measurement data of the corresponding dual-color thermometer is determined to be reliable; otherwise, it is determined to be unreliable. Calibrate the two-color thermometers that are deemed unreliable.

3. The intelligent control system for a quartz crucible melting machine according to claim 2, characterized in that: The process of establishing the arc length-heating temperature correlation model is as follows: Retrieve arc images and synchronously recorded melt temperature data collected from multiple historical melting cycles from the database; The arc image is processed to extract the arc length; The extracted arc length is used as the input variable, and the melt temperature recorded at the corresponding moment is used as the output variable. Regression analysis was used to fit the input and output variables, establish the mapping relationship between arc length and heating temperature, and construct an arc length-heating temperature correlation model.

4. The intelligent control system for a quartz crucible melting machine according to claim 2, characterized in that: The process of establishing the arc power-heating temperature correlation model is as follows: Retrieve arc current and voltage data collected from multiple historical melting cycles and synchronously recorded melt temperature data from the database; The arc power is calculated based on the current and voltage, and the obtained power data is preprocessed. The pre-processed arc power is used as the input variable, and the melt temperature recorded at the corresponding time is used as the output variable. Regression analysis was used to fit the input and output variables, establish the mapping relationship between arc power and heating temperature, and construct an arc power-heating temperature correlation model.

5. The intelligent control system for a quartz crucible melting machine according to claim 1, characterized in that: The specific working process of the preheating module is as follows: During the preheating stage, several testing points are selected on the edge line formed by the contact between the horizontal surface layer of quartz sand and the inner wall of the graphite mold; The temperature at each detection point is collected in real time, and the average value is calculated as the surface temperature of the quartz sand. The heating rate of the surface temperature is monitored and compared with the set expected heating rate to obtain the heating rate deviation. If the current heating rate deviation exceeds the preset allowable range, and the deviation at the next moment still exceeds the range, then it is determined that the heating rate needs to be adjusted. Based on the heating rate deviation, the adjustment direction and amount of the heating rate are determined and adaptive control is performed.

6. The intelligent control system for a quartz crucible melting machine according to claim 5, characterized in that: The specific process for determining the adjustment direction and amount of the heating rate is as follows: The adjustment direction is determined based on the sign of the temperature rise rate deviation; If the deviation is positive, it indicates that the current heating rate is greater than the expected heating rate, and the direction of adjustment of the heating rate is to decrease it; If the deviation is negative, it indicates that the current heating rate is less than the expected heating rate, and the direction of adjustment of the heating rate is to increase it; The absolute value of the heating rate deviation is used as the adjustment amount of the heating rate.

7. The intelligent control system for a quartz crucible melting machine according to claim 1, characterized in that: The specific working process for determining the uniformity of melting in the melting homogenization module is as follows: Using the center point of the melt surface as the origin, concentric rings are set up with a fixed radius step size, and rays are drawn from the origin with a fixed angle step size. Temperature measuring points are set up at the intersection of the concentric rings and the rays. Real-time temperature monitoring of the center point of the melt surface and various temperature measurement points; Calculate the maximum temperature difference between all temperature measuring points on the same ring, and use it as the relative temperature difference on the ring. Calculate the temperature change between temperature measurement points at different radii along the same ray, and construct the radial temperature gradient sequence of the ray; If the following conditions are met simultaneously, the melt is considered to be heated uniformly: (1) The relative temperature difference on all the rings is less than the set relative temperature difference threshold on the rings; (2) The matching degree between the radial temperature gradient sequence of all rays and the reference radial temperature gradient sequence is greater than the set matching degree threshold; Otherwise, it is judged as uneven heating.

8. The intelligent control system for a quartz crucible melting machine according to claim 7, characterized in that: The specific process for identifying the problem type when there is uneven heating in the melting and homogenizing module is as follows: Calculate the average temperature of the ring by averaging the temperatures at each measuring point on the same ring. The average temperature of each ring is compared with the temperature at the center point of the melt surface to obtain the temperature difference. If the temperature difference of all the rings exceeds their respective preset allowable range, it is determined to be an overall mode problem; otherwise, proceed with the following steps: Number the temperature measurement points on the same ring in a set order, establish a coordinate system with the number as the horizontal axis and the temperature as the vertical axis, and draw a scatter plot of the temperature distribution of each ring. The fluctuation bands of data points in the scatter plot are obtained by regression and fitting methods. If a data point deviates from the fluctuation band and its temperature is higher than the upper limit of the fluctuation band, the point is recorded as a local hot spot. If its temperature is lower than the lower limit of the fluctuation band, it is recorded as a local cold spot. Count the number, location, and offset of local hot spots and local cold spots on all the rings.

9. The intelligent control system for a quartz crucible melting machine according to claim 1, characterized in that: The specific working process of the clarification and constant temperature module is as follows: During the clarification stage, the melt temperature is monitored in real time, and its temperature curve over time is plotted. The linear regression line of the temperature curve is used as the baseline. Identify each extreme point on the temperature curve, calculate the fluctuation of each extreme point relative to the baseline, mark the extreme points with fluctuations greater than a set threshold as fluctuation points, and record the time span corresponding to each fluctuation point as the fluctuation duration. If the total number of fluctuation points and the cumulative duration of fluctuations do not exceed their respective set allowable ranges, the temperature fluctuation is determined to be within the limits; otherwise, it is determined to be outside the limits, and the following steps are executed: Based on the morphological characteristics of the temperature curve, the specific pattern of temperature fluctuation is determined. By combining the pre-established association set of each fluctuation pattern with potential causes, the suspected cause of the current temperature fluctuation is obtained and feedback is provided.

10. The intelligent control system for a quartz crucible melting machine according to claim 9, characterized in that: The specific process for determining the temperature fluctuation pattern is as follows: Extract example temperature curves corresponding to each temperature fluctuation pattern pre-stored in the database. The temperature fluctuation patterns include regular oscillations, irregular jumps, unidirectional drifts, and step-like jumps. The temperature curves obtained from real-time monitoring during the clarification phase were compared with the temperature curves of each example based on their shape similarity. The temperature fluctuation pattern corresponding to the example curve with the highest similarity is determined as the final identified fluctuation pattern.

Images