A single crystal furnace thermal field insulation effect detection system based on thermal image analysis

Abstract

The present invention provides a single crystal furnace thermal field insulation effect detection system based on thermal image analysis. Compared with the prior art, the present invention also includes a first thermal imager for monitoring the thermal distribution of the furnace side wall and generating thermal images, a second thermal imager for monitoring the thermal distribution of the furnace bottom and generating thermal images, a third thermal imager for monitoring the thermal distribution of the furnace top and generating thermal images, a cooling module for at least high temperature protection of the second thermal imager, a processing module for processing the image of the thermal distribution monitoring and imaging of the thermal imager, an evaluation module for evaluating the thermal field insulation effect of the single crystal furnace based on the image processed by the processing module, and an alarm module for triggering an alarm when the evaluation module determines that the thermal insulation effect of the single crystal furnace is lower than the expected thermal insulation effect. The present invention adopts multi-parameter comprehensive analysis and dynamic adaptive adjustment to greatly improve the accuracy, flexibility and industrial application value of monitoring the thermal field insulation performance of the single crystal furnace.

Description

Technical Field

[0001] The invention relates to the field of thermal insulation effect detection of a single crystal furnace, and in particular to a thermal insulation effect detection system of a single crystal furnace based on thermal image analysis. Background Art

[0002] In the process of growing crystals by physical vapor transport, crystals are usually grown in a cylindrical crucible by resistance heating or induction heating and cooled to form an ingot. The conventional method of furnace annealing is to gradually reduce the peripheral heating power to slowly cool the ingot temperature to room temperature. However, during the annealing process, the newly condensed crystals mainly show plastic deformation due to the violent molecular movement under high temperature conditions, and no stress is accumulated. But as the temperature gradually decreases, the crystal changes from plastic deformation to elastic deformation and begins to accumulate stress. When the stress exceeds the tensile strength of the crystal or is subjected to external force, the ingot is very easy to crack, resulting in a decrease in material quality. Therefore, the insulation design of the crystal furnace is crucial in the crystal growth process and is an indispensable part of ensuring crystal quality and process stability.

[0003] This experimental team has been browsing and studying a large amount of relevant records and materials on the relevant technologies of the thermal field of single crystal furnaces for a long time. At the same time, relying on relevant resources and conducting a large number of relevant experiments, it has found the existing prior arts such as CN117305969A, CN110284186A, CN219951292U, and CN100371506C disclosed in the prior arts after a large amount of searching. For example, a Czochralski single crystal furnace and a method for measuring and controlling the longitudinal temperature gradient thereof disclosed in the prior art belong to the field of single crystal silicon production. The Czochralski single crystal furnace comprises a crucible and a heater. The periphery of the heater is provided with an insulation layer. The crucible comprises a quartz crucible for holding a melt and an outer supporting crucible sleeved outside the quartz crucible; the heater comprises a main heater surrounding the crucible and an auxiliary heater arranged at the bottom of the crucible; a plurality of thermocouple temperature measuring probes are arranged at intervals in the vertical direction outside the insulation layer, or a first window for infrared thermometer temperature measurement is opened on the insulation layer. By detecting and controlling the temperature distribution of the melt in the quartz crucible, scientific and precise control of the crystal pulling process can be achieved. During the crystal pulling process, the melt temperature at the bottom of the crucible is controlled to be as low as possible to obtain single crystals with lower oxygen content. The maximum temperature of the melt in the crucible is controlled to be as low as possible to obtain single crystals with lower defect density. This not only makes the pulling speed faster and the crystal pulling cost lower, but also greatly improves the quality of the single crystal.

[0004] The present invention is made in order to solve the common problems in the art such as low accuracy of thermal insulation evaluation of single crystal furnaces. Summary of the invention

[0005] The purpose of the present invention is to propose a single crystal furnace thermal field insulation effect detection system based on thermal image analysis in view of the shortcomings existing in the current field.

[0006] In order to overcome the shortcomings of the prior art, the present invention adopts the following technical solutions:

[0007] A single crystal furnace thermal field insulation effect detection system based on thermal image analysis, the single crystal furnace thermal field includes a furnace body, a crucible arranged in the center of the furnace body, a heater arranged in the furnace body to heat the crucible, a heat preservation member covering the heater and the outer periphery of the furnace body, and a support base supporting the bottom of the furnace body.

[0008] The support base includes an isolation metal plate located at the upper part and a railing bracket composed of a plurality of vertical rods located at the lower part. The isolation metal plate is used to directly support the furnace body to isolate the high temperature, and the railing bracket is used to provide a stable support and form a gap.

[0009] The single crystal furnace thermal field insulation effect detection system based on thermal image analysis includes a first thermal imager for monitoring the thermal distribution of the side wall of the furnace body and generating thermal images, a second thermal imager for monitoring the thermal distribution of the bottom of the furnace body and generating thermal images, a third thermal imager for monitoring the thermal distribution of the top of the furnace body and generating thermal images, a cooling module for at least high temperature protection of the second thermal imager, a processing module for processing the images of the thermal imager thermal distribution monitoring and imaging, an evaluation module for evaluating the thermal field insulation effect of the single crystal furnace based on the images processed by the processing module, and an alarm module for triggering an alarm when the evaluation module determines that the insulation effect of the single crystal furnace is lower than the expected insulation effect.

[0010] Furthermore, the second thermal imager is arranged in the gap of the railing bracket, and the cooling module includes a cooling fan and / or a liquid cooler. When the temperature of the operating environment of the second thermal imager is greater than the upper limit operating temperature of the second thermal imager, the cooling module starts to operate to provide cooling protection for the second thermal imager.

[0011] Furthermore, the processing module implements the following steps:

[0012] S101: receiving thermal image data collected by the first, second and third thermal imagers in real time, and adding a source tag and a corresponding single crystal furnace status tag to each frame of the image. The source includes the top, side wall or bottom of the furnace body. The single crystal furnace status includes the heating stage, the insulation stage and the cooling stage. The top, side wall and bottom areas of the furnace body of the source tag are represented by serial numbers 1, 2 and 3 respectively.

[0013] S102: Perform noise filtering on the received thermal image data, apply median filtering or Gaussian filtering to remove random noise, perform contrast enhancement and edge sharpening on the image to highlight the temperature distribution details, and generate an optimized thermal image.

[0014] S103: Converting the grayscale value or pseudo-color value into the actual temperature, extracting the temperature value from the pixel data of the thermal image, and marking and storing the extracted temperature value according to the pixel position of the thermal image to form a temperature data set for each frame of the image.

[0015] S104: According to the preset temperature upper and lower limits, the temperature data is divided into a high temperature area, a low temperature area and a normal temperature area, and a temperature distribution diagram of the furnace body is constructed, wherein the temperature upper limit is 120% of the target temperature of the current state of the furnace body, and the temperature lower limit is 80% of the target temperature of the current state of the furnace body.

[0016] S105: Calculate and output parameter data sets at different positions of the furnace body, wherein the temperature data sets include a temperature mean value, a temperature difference value, a warning value, and an abnormality value.

[0017] S106: Convert the processed temperature data set, temperature distribution map and parameter data into a standard format, add timestamp, source and status information, and save the formatted data to a storage module for subsequent data query and long-term trend analysis. The standard format includes a table, a CSV file and / or an image sequence.

[0018] S107: The processed temperature data set, temperature distribution diagram and parameter data are transmitted to the evaluation module to provide complete input data for the evaluation of the thermal insulation effect.

[0019] Furthermore, the specific calculation steps of the temperature data in step S105 are as follows:

[0020] S1051: Obtain a thermal image from a thermal imager, convert the grayscale value of each pixel in the thermal image into an actual temperature value, and use T1 （x，y） T2 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 1 collected by the first thermal imager. （x，y） T3 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 2 collected by the second thermal imager. （x，y） It is represented by the actual temperature value of the pixel at the position (x, y) in the thermal image of the 3-numbered area collected by the third thermal imager, and Ti is represented by the thermal image of the i-numbered area, i={1,2,3}, Ti （x，y） Represents the actual temperature value of the pixel at position (x, y) in the thermal image of region i.

[0021] S1052: Calculate the temperature mean Ti of the 1st, 2nd and 3rd serial number zones respectively mean :

[0022] ,

[0023] Among them, Ti meanis the mean temperature of the i-sequence area, ni and mi are the total number of rows and columns of the i-sequence area, Ni is the total number of pixels in the i-sequence area, ni×mi=Ni,

[0024] S1053: Calculate the temperature difference Ti of the 1st, 2nd and 3rd serial number zones respectively var :

[0025] , where Ti var is the temperature difference value of the i-numbered zone,

[0026] S1054: Calculate the warning values ​​Pi for the 1st, 2nd and 3rd sequence number areas respectively alert : ,

[0027] Among them, Z is the classification parameter, when Ti （x，y） ≥Ti thr When Z=1, when Ti （x，y） <Ti thr When Z=0, Pi alert is the warning value of the i-th serial number zone, Ti thr is the temperature threshold of zone i. The warning value is greater than the proportion of temperature exceeding the standard point, which reflects whether the local temperature is within the safe range and is a manifestation of the stability of the temperature control system.

[0028] The warning value is inversely proportional to the insulation effect of the insulation component. The better the performance of the insulation component, the smaller the heat loss, the lower the proportion of temperature abnormalities, and the lower the warning value. On the contrary, when the performance of the insulation component is poor, the temperature control difficulty increases, the warning value rises, indicating that the insulation effect is poor.

[0029] S1055: Calculate the temperature gradient values ​​TNi of the numbered zones 1, 2, and 3 respectively:

[0030] ,

[0031] △x is the pixel spacing distance in the horizontal direction, △y is the adjacent pixel spacing distance in the vertical direction, and △x and △y are a unit pixel. The temperature gradient value mainly measures the uniformity of the temperature field and reflects the severity of the temperature change in different areas of the furnace body. It is a manifestation of the performance of the insulation parts and the stability of the thermal field. The higher the temperature gradient value, the uneven distribution of the furnace temperature field and the larger the local temperature difference, which corresponds to the damage of the insulation material or the decrease of the insulation performance. On the contrary, the lower the temperature gradient value, the more uniform the distribution of the furnace temperature field, the smaller the heat loss, and the better the insulation effect of the insulation parts.

[0032] Furthermore, the evaluation module implements the following operation steps:

[0033] S201: receiving the warning value P1 of the thermal image collected by the first, second and third thermal imagers at the same time respectively calculated by the processing module alert 、P2 alert 、P3 alert , and identify the single crystal furnace state corresponding to the picture, and determine the standard warning value Pth of the corresponding single crystal furnace state,

[0034] exist When the temperature of the furnace is within the normal range,

[0035] exist When further performing the following steps S202-S203,

[0036] Among them, a specific warning standard value is pre-set for each single crystal furnace state, and the warning standard values ​​in the heating stage, insulation stage and cooling stage are expressed as Pj, Pb, Pr, Pth∈{Pj, Pb, Pr}, Pj≥Pb≥Pr,

[0037] S202: Calculate the comprehensive temperature distribution index of the furnace body: ,

[0038] Among them, Max (Ti var ) are all T1 calculated in the thermal image data var 、T2 var 、T3 var The maximum value in which Max (Pi alert ) are all P1 calculated in the thermal image data alert 、P2 alert 、P3 alert Max(TNi) is the maximum value among all TN1, TN2, and TN3 calculated in the thermal image data, RE1 is a temperature difference value standardization parameter used to standardize the temperature difference value so that it is comparable in calculation, RE2 is a warning value standardization parameter used to standardize the expected value so that it is comparable in calculation, RE3 is a temperature gradient value standardization parameter used to standardize the temperature gradient value so that it is comparable in calculation, α is the temperature difference weight coefficient, β is the warning value weight coefficient, λ is the temperature gradient weight coefficient,

[0039] S203: Calculate the comprehensive insulation effect index BI of the single crystal furnace under different conditions:

[0040] ;

[0041] RF1 is the target operating temperature of the single crystal furnace in the cooling state, RF2 is the target operating temperature of the single crystal furnace in the insulation state, and RF3 is the target operating temperature of the single crystal furnace in the heating state. The target operating temperature values ​​corresponding to different states reflect the ideal temperature benchmark of the single crystal furnace when it is running at each stage and serve as a reference for evaluating the actual insulation effect. w1 and w2 are weight coefficients, and w1≤w2 is set.

[0042] S204: When BI is greater than 0, it is judged that the thermal insulation effect of the single crystal furnace is lower than the expected thermal insulation effect, and it is predicted that the single crystal furnace is damaged.

[0043] On the contrary, when BI is less than or equal to 0, it is judged that the thermal insulation effect of the single crystal furnace is above the expected thermal insulation effect, and the single crystal furnace is judged to be normal.

[0044] Furthermore, when the evaluation module determines that the insulation effect of the single crystal furnace is lower than the expected insulation effect, the alarm module pushes the alarm information through SMS, email or monitoring platform to notify the on-site operators and system administrators, and stores all the alarm data to facilitate the subsequent fault analysis of the single crystal furnace by the on-site operators and system administrators.

[0045] The alarm data includes the occurrence time, the position of the single crystal furnace, the state of the single crystal furnace, the latest parameter data set of the single crystal furnace, the latest temperature data set of the single crystal furnace and the latest temperature distribution diagram of the single crystal furnace.

[0046] The beneficial effects achieved by the present invention are:

[0047] 1. The present invention covers the top, side walls and bottom of the furnace body with multiple thermal imagers, combined with efficient data processing modules, to monitor temperature distribution and anomalies in real time, and accurately calculate key indicators such as temperature mean, difference, gradient, etc. It dynamically adjusts the analysis strategy according to the heating, insulation and cooling states of the furnace body to ensure multi-state adaptability, and at the same time standardizes and saves the processed data to support long-term trend analysis. The overall design enhances the accuracy and comprehensiveness of temperature monitoring, and provides reliable guarantee for stable operation of equipment and thermal field optimization.

[0048] 2. The present invention can not only monitor and evaluate the insulation effect of the single crystal furnace thermal field in real time, but also ensure the temperature stability and distribution uniformity of the furnace body under different conditions through comprehensive data processing and fine indicator analysis. In particular, through dynamic monitoring of temperature gradients and early warning values, it helps to timely discover and solve potential insulation problems, improve equipment operation efficiency and product quality, and provide technical support for the intelligent upgrade of industrial production processes.

[0049] 3. The evaluation module introduces multi-dimensional parameters such as temperature difference, warning value and temperature gradient, and ensures the comprehensiveness and accuracy of the evaluation results through standardization and weighted calculation. It designs adaptive evaluation strategies for the three states of heating, insulation and cooling, focusing on temperature uniformity, local over-temperature and heat dissipation stability, respectively, to improve the pertinence of temperature control management, and uses a neural network model to dynamically generate empirical parameters. Through continuous optimization and regular updating, the accuracy, adaptability and long-term effectiveness of the evaluation are enhanced.

[0050] 4. The present invention accurately detects the thermal insulation effect of the single crystal furnace through multi-dimensional comprehensive parameter evaluation, state adaptive adjustment and neural network optimization, and realizes dynamic adaptation and long-term reliability of temperature control requirements at each stage. At the same time, the efficient data processing and storage mechanism of the present invention provides a basis for long-term analysis and optimization of the thermal insulation effect, ensuring that the actual application value of the evaluation system in industrial production is significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0052] Figure 1 It is a modular schematic diagram of the single crystal furnace thermal field insulation effect detection system based on thermal image analysis of the present invention.

[0053] Figure 2 It is a schematic diagram of experimental data between the comprehensive insulation effect index BI of the present invention and the thermal insulation effect of the single crystal furnace thermal field. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with its embodiments; it should be pointed out that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present case. For those skilled in the art, after reviewing the following detailed description, other systems, methods and / or features of this embodiment will become apparent. And the terms describing the positional relationship in the accompanying drawings are only used for illustrative purposes and cannot be understood as limitations on the present application. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0055] Embodiment 1: Combining with Figure 1 and attached Figure 2The present embodiment constructs a single crystal furnace thermal field insulation effect detection system based on thermal image analysis, wherein the single crystal furnace thermal field includes a furnace body, a crucible arranged at the center of the furnace body, a heater arranged in the furnace body for heating the crucible, a heat preservation member covering the heater and the periphery of the furnace body, and a support base for supporting the bottom of the furnace body, wherein the support base includes an isolation metal plate correspondingly located at the upper part and a railing bracket composed of a plurality of vertical rods correspondingly located at the lower part, the isolation metal plate is used to directly support the furnace body to isolate the high temperature, and the railing bracket is used to provide stable support and form a gap.

[0056] The single crystal furnace thermal field insulation effect detection system based on thermal image analysis includes a first thermal imager for monitoring the thermal distribution of the side wall of the furnace body and generating thermal images, a second thermal imager for monitoring the thermal distribution of the bottom of the furnace body and generating thermal images, a third thermal imager for monitoring the thermal distribution of the top of the furnace body and generating thermal images, a cooling module for at least high temperature protection of the second thermal imager, a processing module for processing the images of thermal distribution monitoring and imaging by the thermal imager, an evaluation module for evaluating the thermal field insulation effect of the single crystal furnace based on the images processed by the processing module, and an alarm module for triggering an alarm when the evaluation module determines that the insulation effect of the single crystal furnace is lower than the expected insulation effect, and the second thermal imager is arranged in the gap of the railing bracket.

[0057] The cooling module is a cooling fan and / or liquid cooler in the prior art. When the temperature of the operating environment of the second thermal imager is greater than the upper limit operating temperature of the second thermal imager, the cooling module starts to operate to cool down and protect the second thermal imager.

[0058] The processing module implements the following steps:

[0059] S101: receiving thermal image data collected by the first, second and third thermal imagers in real time, and adding a source tag and a corresponding single crystal furnace status tag to each frame of the image, the source includes the top, side wall or bottom of the furnace body, the single crystal furnace status includes the heating stage, the insulation stage and the cooling stage, and the top, side wall and bottom areas of the furnace body of the source tag are represented by serial numbers 1, 2 and 3 respectively;

[0060] S102: performing noise filtering on the received thermal image data, applying median filtering or Gaussian filtering to remove random noise, performing contrast enhancement and edge sharpening processing on the image to highlight temperature distribution details, and generating an optimized thermal image;

[0061] S103: extracting temperature values ​​from pixel data of the thermal image by converting grayscale values ​​or pseudo-color values ​​into actual temperatures, and marking and storing the extracted temperature values ​​according to pixel positions of the thermal image to form a temperature data set for each frame of the image;

[0062] S104: According to the preset temperature upper and lower limits, the temperature data is divided into a high temperature area, a low temperature area and a normal temperature area, and a temperature distribution map of the furnace body is constructed, wherein the temperature upper limit is 120% of the target temperature of the current state of the furnace body, and the temperature lower limit is 80% of the target temperature of the current state of the furnace body;

[0063] S105: Calculate and output parameter data sets at different positions of the furnace body, wherein the temperature data sets include a temperature mean value, a temperature difference value, a warning value, and an abnormality degree value;

[0064] The specific calculation steps of the temperature data in step S105 are as follows:

[0065] S1051: Obtain a thermal image from a thermal imager, convert the grayscale value of each pixel in the thermal image into an actual temperature value, and use T1 （x，y） T2 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 1 collected by the first thermal imager. （x，y） T3 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 2 collected by the second thermal imager. （x，y） It is represented by the actual temperature value of the pixel at the position (x, y) in the thermal image of the 3-numbered area collected by the third thermal imager, and Ti is represented by the thermal image of the i-numbered area, i={1,2,3}, Ti （x，y） Represents the actual temperature value of the pixel at position (x, y) in the thermal image of region i.

[0066] S1052: Calculate the temperature mean Ti of the 1st, 2nd and 3rd serial number zones respectively mean :

[0067] ,

[0068] Among them, Ti mean is the mean temperature of the i-sequence area, ni and mi are the total number of rows and columns of the i-sequence area, Ni is the total number of pixels in the i-sequence area, ni×mi=Ni,

[0069] S1053: Calculate the temperature difference Ti of the 1st, 2nd and 3rd serial number zones respectively var :

[0070] , where Ti var is the temperature difference value of the serial number zone i, S1054: calculate the warning value Pi of the serial number zones 1, 2, and 3 respectively alert :

[0071] ,

[0072] Among them, Z is the classification parameter, when Ti（x，y） ≥Ti thr When Z=1, when Ti （x，y） <Ti thr When Z=0, Pi alert is the warning value of the i-th serial number zone, Ti thr is the temperature threshold of zone i. The warning value is greater than the proportion of temperature exceeding the standard point, which reflects whether the local temperature is within the safe range and is a manifestation of the stability of the temperature control system.

[0073] The warning value is inversely proportional to the insulation effect of the insulation component. The better the performance of the insulation component, the smaller the heat loss, the lower the proportion of temperature abnormality points, and the lower the warning value. On the contrary, when the performance of the insulation component is poor, the temperature control difficulty increases and the warning value increases, indicating that the insulation effect is poor.

[0074] S1055: Calculate the temperature gradient values ​​TNi of the numbered zones 1, 2, and 3 respectively:

[0075] ,

[0076] △x is the pixel spacing distance in the horizontal direction, △y is the adjacent pixel spacing distance in the vertical direction, and △x and △y are a unit pixel. The temperature gradient value mainly measures the uniformity of the temperature field, reflects the severity of the temperature change in different areas of the furnace, and is a manifestation of the performance of the insulation component and the stability of the thermal field. The higher the temperature gradient value, the uneven distribution of the furnace temperature field and the larger the local temperature difference, which corresponds to the damage of the insulation component material or the decline of the insulation performance. On the contrary, the lower the temperature gradient value, the more uniform the furnace temperature field distribution, the smaller the heat loss, and the better the insulation effect of the insulation component.

[0077] S106: converting the processed temperature data set, temperature distribution map and parameter data into a standard format, adding timestamp, source and status information, and saving the formatted data to a storage module to facilitate subsequent data query and long-term trend analysis, wherein the standard format includes a table, a CSV file and / or an image sequence;

[0078] S107: The processed temperature data set, temperature distribution diagram and parameter data are transmitted to the evaluation module to provide complete input data for the evaluation of the thermal insulation effect.

[0079] The specific steps of executing the temperature data in S105 are as follows:

[0080] import numpy as np

[0081] # Constants: Define pixel spacings

[0082] delta_x = 1 # Horizontal pixel spacing

[0083] delta_y = 1 # Vertical pixel spacing

[0084] # Input temperature data: A 3D numpy array where T[i] is the (ni, mi)array for region i

[0085] # Replace this with actual data

[0086] T = [np.random.random((10, 10)) * 100, # Region 1 (top)

[0087] np.random.random((10, 10)) * 100, # Region 2 (sidewall)

[0088] np.random.random((10, 10)) * 100] # Region 3 (bottom)

[0089] # S1051: Convert grayscale to actual temperatures (assume data isalready in temperature)

[0090] def calculate_temperature_means(T):

[0091] # S1052: Calculate mean temperature for each region

[0092] T_mean = [np.mean(region) for region in T]

[0093] return T_mean

[0094] # S1053: Calculate temperature variance for each region

[0095] def calculate_temperature_variance(T, T_mean):

[0096] T_var = [np.mean((region - mean) ** 2) for region, mean in zip(T,T_mean)]

[0097] return T_var

[0098] # S1054: Calculate temperature gradient (TNi) for each region

[0099] def calculate_temperature_gradient(T):

[0100] T_ni = []

[0101] for region in T:

[0102] # Gradient calculations (forward difference)

[0103] dT_dx = (region[:, 1:] - region[:, :-1]) / delta_x #Horizontal gradient

[0104] dT_dy = (region[1:, :] - region[:-1, :]) / delta_y # Verticalgradient

[0105] # Combine gradients

[0106] gradient_magnitude = np.sqrt(dT_dx[:, :-1] ** 2 + dT_dy[:-1,:] ** 2)

[0107] # Mean gradient as TNi

[0108] T_ni.append(np.mean(gradient_magnitude))

[0109] return T_ni

[0110] # S1053: Calculate alert parameter for each region

[0111] def calculate_alert_parameters(T, threshold):

[0112] alert_params = []

[0113] for region in T:

[0114] # Count the number of pixels exceeding threshold

[0115] alert_param = np.sum(region >= threshold) / region.size

[0116] alert_params.append(alert_param)

[0117] return alert_params

[0118] # Main computation for S1051-S1054

[0119] if __name__ == "__main__":

[0120] # Step S1052

[0121] T_mean = calculate_temperature_means(T)

[0122] print("Mean Temperatures (T_mean):", T_mean)

[0123] # Step S1053

[0124] T_var = calculate_temperature_variance(T, T_mean)

[0125] print("Temperature Variances (T_var):", T_var)

[0126] # Alert thresholds for each region

[0127] thresholds = [80, 80, 80] # Example thresholds

[0128] P_alert = calculate_alert_parameters(T, thresholds)

[0129] print("Alert Parameters (P_alert):", P_alert)

[0130] # Step S1054

[0131] T_ni = calculate_temperature_gradient(T)

[0132] print("Temperature Gradients (TNi):", T_ni).

[0133] The present invention covers the top, side walls and bottom of the furnace body with multiple thermal imagers, and combines with efficient data processing modules to monitor temperature distribution and anomalies in real time, and accurately calculate key indicators such as temperature mean, difference, gradient, etc. It dynamically adjusts the analysis strategy according to the heating, insulation and cooling states of the furnace body to ensure multi-state adaptability, and at the same time standardizes and saves the processed data to support long-term trend analysis. The overall design enhances the accuracy and comprehensiveness of temperature monitoring, and provides reliable guarantee for stable operation of equipment and thermal field optimization.

[0134] The present invention can not only monitor and evaluate the insulation effect of the single crystal furnace thermal field in real time, but also ensure the temperature stability and distribution uniformity of the furnace body under different conditions through comprehensive data processing and sophisticated indicator analysis. In particular, dynamic monitoring of temperature gradients and warning values ​​is helpful to timely discover and solve potential insulation problems, improve equipment operation efficiency and product quality, and provide technical support for the intelligent upgrade of industrial production processes.

[0135] Embodiment 2: Combining with Figure 1 and attached Figure 2 In addition to the contents of the above embodiments, the evaluation module implements the following operation steps:

[0136] S201: receiving the warning value P1 of the thermal image collected by the first, second and third thermal imagers at the same time respectively calculated by the processing module alert 、P2 alert 、P3 alert , and identify the single crystal furnace state corresponding to the picture, and determine the standard warning value Pth of the corresponding single crystal furnace state,

[0137] exist When the temperature of the furnace is within the normal range,

[0138] exist When further performing the following steps S202-S203,

[0139] Among them, a specific warning standard value is pre-set for each single crystal furnace state, and the warning standard values ​​in the heating stage, insulation stage and cooling stage are expressed as Pj, Pb, Pr, Pth∈{Pj, Pb, Pr}, Pj≥Pb≥Pr. In the heating stage, the warning value is high, allowing local overheating to improve the heating efficiency. In the insulation stage, the warning value is the lowest, and temperature abnormalities are strictly controlled to ensure stability. In the cooling stage, the warning value is moderate, and attention is paid to the risk of local over-cooling or uneven heat dissipation.

[0140] The warning standard values ​​under different conditions of the furnace body are determined by technicians in this field through experimental measurement, historical operation data analysis, numerical simulation and deep learning optimization, so that the warning values ​​can accurately reflect the normal operation range and remain sensitive to abnormal conditions, thereby optimizing the thermal insulation performance evaluation of the furnace body at each stage;

[0141] S202: Calculate the comprehensive temperature distribution index TR of the furnace body: ,

[0142] Among them, Max (Ti var ) are all T1 calculated in the thermal image data var 、T2 var 、T3 var The maximum value in which Max (Pi alert ) are all P1 calculated in the thermal image data alert 、P2 alert 、P3 alert Max (TNi) is the maximum value of all TN1, TN2, and TN3 calculated in the thermal image data, RE1 is a temperature difference value standardization parameter used to standardize the temperature difference value so that it is comparable in calculation, RE2 is a warning value standardization parameter used to standardize the predicted value so that it is comparable in calculation, and RE3 is a temperature gradient value standardization parameter used to standardize the temperature gradient value so that it is comparable in calculation, wherein RE1, RE2, and RE3 are determined and obtained by those skilled in the art based on historical data analysis, working condition simulation, and experimental calibration, and are used to normalize different values ​​so that the influence of different indicators can be reasonably reflected in the calculation.

[0143] α is the temperature difference weight coefficient, β is the warning value weight coefficient, and λ is the temperature gradient weight coefficient. α, β and λ are determined according to the weight ratio of the factors affecting the thermal insulation effect of the furnace body under different states. Among them, α mainly reflects the influence of temperature distribution uniformity on system stability, β reflects the sensitivity of the high temperature over-temperature ratio of the furnace body to safety performance, and λ is used to measure the comprehensive effect of the temperature field gradient change of the furnace body on the thermal insulation performance. α, β and λ are based on historical operation data analysis and experimental calibration results, and combined with actual working conditions to adjust the degree of influence of each indicator on the thermal insulation performance. Through numerical simulation and multi-dimensional optimization algorithm, α, β and λ are rationalized and optimized on the basis of meeting the requirements of different single crystal furnace states, and the final values ​​of α, β and λ are obtained;

[0144] S203: Calculate the comprehensive insulation effect index BI of the single crystal furnace under different conditions:

[0145] ;

[0146] RF1 is the target operating temperature of the single crystal furnace in the cooling state, RF2 is the target operating temperature of the single crystal furnace in the insulation state, and RF3 is the target operating temperature of the single crystal furnace in the heating state. The target operating temperature values ​​corresponding to different states reflect the ideal temperature benchmark of the single crystal furnace when it is running at each stage and serve as a reference for evaluating the actual insulation effect. When BI is greater than 0, it is judged that the insulation effect of the single crystal furnace is lower than the expected insulation effect. When BI is less than or equal to 0, it is judged that the insulation effect of the single crystal furnace is higher than the expected insulation effect. w1 and w2 are weight coefficients, and w1≤w2 is set;

[0147] In the cooling state, the maximum temperature and unevenness are amplified through square processing to judge overheating and unevenness in the cooling process. In the insulation state, the unevenness is smoothed logarithmically to avoid excessive response to small fluctuations and judge the overall stability of the constant temperature stage. In the heating state, the temperature extremes are paid attention to to ensure uniform temperature distribution during the heating process, so as to achieve targeted evaluation of the insulation effect of the furnace under different states, so that accurate temperature monitoring and analysis can be achieved in each stage.

[0148] The ex1, ex2, and ex3 are respectively used to adapt to the expected insulation requirements under different states of the single crystal furnace, and their values ​​are derived from the following process: technicians in this field collect and preprocess a large number of parameter data sets, and mark the corresponding furnace operation states and insulation effects to build a neural network model. The model uses one or more of a multilayer perceptron, a convolutional neural network, and a long short-term memory network. During the model training process, the parameter data set is used as input, and the error between the insulation effect evaluation index BI and the actual evaluation value is minimized by defining a suitable loss function optimization. After repeated training and verification, the model gradually converges and generates the optimal empirical parameters ex1, ex2, and ex3 to meet the insulation requirements in each state.

[0149] Finally, the verified and tuned neural network model can generate reasonable empirical parameters ex1, ex2, and ex3 according to different states, making the evaluation of thermal insulation effect more accurate and adaptable. At the same time, in order to ensure the long-term reliability and evaluation accuracy of the model, the neural network can be regularly retrained and parameter updated as new data accumulates, so as to continuously optimize the values ​​of ex1, ex2, and ex3 and enhance their applicability.

[0150] The evaluation module introduces multi-dimensional parameters such as temperature difference, warning value and temperature gradient, and ensures the comprehensiveness and accuracy of the evaluation results through standardization and weighted calculation. It designs adaptive evaluation strategies for the three states of heating, insulation and cooling, focusing on temperature uniformity, local over-temperature and heat dissipation stability, respectively, to improve the targeted temperature control management, and uses a neural network model to dynamically generate empirical parameters. Through continuous optimization and regular updating, the accuracy, adaptability and long-term effectiveness of the evaluation are enhanced.

[0151] When the evaluation module determines that the insulation effect of the single crystal furnace is lower than the expected insulation effect, the alarm module pushes the alarm information via SMS, email or monitoring platform to notify the on-site operators and system administrators, and stores all alarm data to facilitate subsequent fault analysis of the single crystal furnace by the on-site operators and system administrators.

[0152] The alarm data includes the occurrence time, the position of the single crystal furnace, the state of the single crystal furnace, the latest parameter data set of the single crystal furnace, the latest temperature data set of the single crystal furnace and the latest temperature distribution diagram of the single crystal furnace.

[0153] The present invention accurately detects the thermal insulation effect of the single crystal furnace thermal field through multi-dimensional comprehensive parameter evaluation, state adaptive adjustment and neural network optimization, and realizes dynamic adaptation and long-term reliability of temperature control requirements at each stage. At the same time, the efficient data processing and storage mechanism of the present invention provides a basis for long-term analysis and optimization of the thermal insulation effect, ensuring that the actual application value of the evaluation system in industrial production is significantly improved.

[0154] Although the present invention has been described above with reference to various embodiments, it should be understood that many changes and modifications may be made without departing from the scope of the present invention. That is, the methods, systems and devices discussed above are examples. Various configurations may appropriately omit, replace or add various processes or components. For example, in alternative configurations, the method may be performed in an order different from that described, and / or various components may be added, omitted and / or combined. Moreover, the features described with respect to certain configurations may be combined in various other configurations, such as different aspects and elements of the configuration may be combined in a similar manner. In addition, as technology develops, the elements therein may be updated, i.e., many elements are examples and do not limit the scope of the present disclosure or claims. And it should be understood that after reading the contents of the record of the present invention, the technician may make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. A single crystal furnace thermal field insulation effect detection system based on thermal image analysis, wherein the single crystal furnace thermal field includes a furnace body, a crucible arranged at the center of the furnace body, a heater arranged in the furnace body to heat the crucible, a heat preservation member covering the heater and the outer periphery of the furnace body, and a support base supporting the bottom of the furnace body. The support base includes an isolation metal plate located at the upper part and a railing bracket composed of a plurality of vertical rods located at the lower part. The isolation metal plate is used to directly support the furnace body to isolate the high temperature, and the railing bracket is used to provide a stable support and form a gap. It is characterized in that The thermal field insulation effect detection system of the single crystal furnace based on thermal image analysis includes a first thermal imager for monitoring the thermal distribution of the side wall of the furnace body and generating a thermal image, a second thermal imager for monitoring the thermal distribution of the bottom of the furnace body and generating a thermal image, a third thermal imager for monitoring the thermal distribution of the top of the furnace body and generating a thermal image, a cooling module for at least performing high temperature protection on the second thermal imager, a processing module for processing the images of the thermal distribution monitoring and imaging of the thermal imager, an evaluation module for evaluating the thermal field insulation effect of the single crystal furnace based on the images processed by the processing module, and an alarm module for triggering an alarm when the evaluation module determines that the thermal insulation effect of the single crystal furnace is lower than the expected thermal insulation effect; The second thermal imager is arranged at the gap of the railing bracket, and the cooling module includes a cooling fan and / or a liquid cooler. When the temperature of the operating environment of the second thermal imager is greater than the upper limit operating temperature of the second thermal imager, the cooling module starts to operate to cool down and protect the second thermal imager; The processing module implements the following steps: S101: receiving thermal image data collected by the first, second and third thermal imagers in real time, and adding a source tag and a corresponding single crystal furnace status tag to each frame of the image. The source includes the top, side wall or bottom of the furnace body. The single crystal furnace status includes the heating stage, the insulation stage and the cooling stage. The top, side wall and bottom areas of the furnace body of the source tag are represented by serial numbers 1, 2 and 3 respectively. S102: Perform noise filtering on the received thermal image data, apply median filtering or Gaussian filtering to remove random noise, perform contrast enhancement and edge sharpening on the image to highlight the temperature distribution details, and generate an optimized thermal image. S103: Converting the grayscale value or pseudo-color value into the actual temperature, extracting the temperature value from the pixel data of the thermal image, and marking and storing the extracted temperature value according to the pixel position of the thermal image to form a temperature data set for each frame of the image. S104: According to the preset temperature upper and lower limits, the temperature data is divided into a high temperature area, a low temperature area and a normal temperature area, and a temperature distribution diagram of the furnace body is constructed, wherein the temperature upper limit is 120% of the target temperature of the current state of the furnace body, and the temperature lower limit is 80% of the target temperature of the current state of the furnace body. S105: Calculate and output parameter data sets at different positions of the furnace body, wherein the temperature data sets include a temperature mean value, a temperature difference value, a warning value, and an abnormality value. S106: Convert the processed temperature data set, temperature distribution map and parameter data into a standard format, add timestamp, source and status information, and save the formatted data to a storage module for subsequent data query and long-term trend analysis. The standard format includes a table, a CSV file and / or an image sequence. S107: transmitting the processed temperature data set, temperature distribution diagram and parameter data to the evaluation module to provide complete input data for the evaluation of the thermal insulation effect; The specific calculation steps of the temperature data in step S105 are as follows: S1051: Obtain a thermal image from a thermal imager, convert the grayscale value of each pixel in the thermal image into an actual temperature value, and use T1 （x，y） T2 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 1 collected by the first thermal imager. （x，y） T3 represents the actual temperature value of the pixel at position (x, y) in the thermal image of area 2 collected by the second thermal imager. （x，y） It is represented by the actual temperature value of the pixel at the position (x, y) in the thermal image of the 3-numbered area collected by the third thermal imager, and Ti is represented by the thermal image of the i-numbered area, i={1,2,3}, Ti （x，y） Represents the actual temperature value of the pixel at position (x, y) in the thermal image of region i. S1052: Calculate the temperature mean Ti of the 1st, 2nd and 3rd serial number zones respectively mean : ， Among them, Ti mean is the mean temperature of the i-sequence area, ni and mi are the total number of rows and columns of the i-sequence area, Ni is the total number of pixels in the i-sequence area, ni×mi=Ni, S1053: Calculate the temperature difference Ti of the 1st, 2nd and 3rd serial number zones respectively var : , where Ti var is the temperature difference value of the i-numbered zone, S1054: Calculate the warning values ​​Pi for the 1st, 2nd and 3rd sequence number areas respectively alert : ， Among them, Z is the classification parameter, when Ti （x，y） ≥Ti thr When Z=1, when Ti （x，y） <Ti thr When Z=0, Pi alert is the warning value of the i-th serial number zone, Ti thr is the temperature threshold of zone i. The warning value is greater than the proportion of temperature exceeding the standard point, which reflects whether the local temperature is within the safe range and is a manifestation of the stability of the temperature control system. The warning value is inversely proportional to the insulation effect of the insulation component. The better the performance of the insulation component, the smaller the heat loss, the lower the proportion of temperature abnormalities, and the lower the warning value. On the contrary, when the performance of the insulation component is poor, the temperature control difficulty increases, the warning value rises, indicating that the insulation effect is poor. S1055: Calculate the temperature gradient values ​​TNi of the numbered zones 1, 2, and 3 respectively: ， △x is the pixel spacing distance in the horizontal direction, △y is the adjacent pixel spacing distance in the vertical direction, and △x and △y are a unit pixel. The temperature gradient value mainly measures the uniformity of the temperature field, reflects the severity of the temperature change in different areas of the furnace, and is a manifestation of the performance of the insulation component and the stability of the thermal field. The higher the temperature gradient value, the uneven distribution of the furnace temperature field and the larger the local temperature difference, which corresponds to the damage of the insulation component material or the decline of the insulation performance. On the contrary, the lower the temperature gradient value, the more uniform the furnace temperature field distribution, the smaller the heat loss, and the better the insulation effect of the insulation component. The evaluation module implements the following operating steps: S201: receiving the warning value P1 of the thermal image collected by the first, second and third thermal imagers at the same time respectively calculated by the processing module alert 、P2 alert 、P3 alert , and identify the single crystal furnace state corresponding to the picture, and determine the standard warning value Pth of the corresponding single crystal furnace state, exist When the temperature of the furnace is within the normal range, exist When further performing the following steps S202-S203, Among them, a specific warning standard value is pre-set for each single crystal furnace state, and the warning standard values ​​in the heating stage, insulation stage and cooling stage are expressed as Pj, Pb, Pr, Pth∈{Pj, Pb, Pr}, Pj≥Pb≥Pr, S202: Calculate the comprehensive temperature distribution index TR of the furnace body: ， Among them, Max (Ti var ) are all T1 calculated in the thermal image data var 、T2 var 、T3 var The maximum value in which Max (Pi alert ) are all P1 calculated in the thermal image data alert 、P2 alert 、P3 alert Max(TNi) is the maximum value among all TN1, TN2, and TN3 calculated in the thermal image data, RE1 is a temperature difference value standardization parameter used to standardize the temperature difference value so that it is comparable in calculation, RE2 is a warning value standardization parameter used to standardize the expected value so that it is comparable in calculation, RE3 is a temperature gradient value standardization parameter used to standardize the temperature gradient value so that it is comparable in calculation, α is the temperature difference weight coefficient, β is the warning value weight coefficient, λ is the temperature gradient weight coefficient, S203: Calculate the comprehensive insulation effect index BI of the single crystal furnace under different conditions: ； RF1 is the target operating temperature of the single crystal furnace in the cooling state, RF2 is the target operating temperature of the single crystal furnace in the insulation state, and RF3 is the target operating temperature of the single crystal furnace in the heating state. The target operating temperature values ​​corresponding to different states reflect the ideal temperature benchmark of the single crystal furnace when it is running at each stage and serve as a reference for evaluating the actual insulation effect. w1 and w2 are weight coefficients, and w1≤w2 is set. Among them, ex1, ex2, and ex3 are respectively used to adapt to the expected insulation requirements of the single crystal furnace under different conditions, and their values ​​are derived from the following process: by collecting and preprocessing a large number of parameter data sets, and marking the corresponding furnace operation status and insulation effect, a neural network model is constructed. The model uses one or more of a multilayer perceptron, a convolutional neural network, and a long short-term memory network. During the model training process, the parameter data set is used as input, and the error between the insulation effect evaluation index BI and the actual evaluation value is minimized by defining a suitable loss function optimization. After repeated training and verification, the model gradually converges and generates the optimal empirical parameters ex1, ex2, and ex3 to meet the insulation requirements under each state; S204: When BI is greater than 0, it is judged that the thermal insulation effect of the single crystal furnace is lower than the expected thermal insulation effect, and it is predicted that the single crystal furnace is damaged. On the contrary, when BI is less than or equal to 0, it is judged that the thermal insulation effect of the single crystal furnace is above the expected thermal insulation effect, and the single crystal furnace is judged to be normal.

2. The single crystal furnace thermal field insulation effect detection system according to claim 1, characterized in that: When the evaluation module determines that the insulation effect of the single crystal furnace is lower than the expected insulation effect, the alarm module pushes the alarm information through SMS, email or monitoring platform to notify the on-site operators and system administrators, and stores all the alarm data to facilitate the subsequent fault analysis of the single crystal furnace by the on-site operators and system administrators. The alarm data includes the occurrence time, the position of the single crystal furnace, the state of the single crystal furnace, the latest parameter data set of the single crystal furnace, the latest temperature data set of the single crystal furnace and the latest temperature distribution diagram of the single crystal furnace.

Images