Temperature control optimization method, system and equipment for semiconductor magnetic core preparation

Abstract

The invention provides a temperature control optimization method, system and equipment for semiconductor magnetic core preparation, and relates to the field of semiconductor magnetic core preparation, and the method comprises the steps: configuring a temperature monitoring array according to a temperature partitioning result of a high temperature furnace; in the semiconductor magnetic core sintering process, regular monitoring is carried out through a temperature monitoring array, and a plurality of temperature data sequences are obtained; multiple pieces of magnetic core sintering information are obtained through prediction according to the multiple temperature data sequences; and an annealing temperature control scheme is configured according to the multiple pieces of magnetic core sintering information, and annealing temperature zoning control is carried out based on a temperature zoning result. The technical problems that in the traditional semiconductor magnetic core preparation process, annealing control accuracy is insufficient, and temperature control response lags behind are solved. Through accurate temperature monitoring and zoning control, the temperature uniformity and stability in the annealing process can be improved, so that the performance and consistency of the magnetic core are optimized, the defect rate is reduced, and the production efficiency and the energy efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor magnetic core preparation, and in particular to a temperature control optimization method, system and equipment for semiconductor magnetic core preparation. Background Art

[0002] Semiconductor cores are widely used in electronic devices, especially in the fields of power, electromagnetic compatibility (EMC) and signal processing. The preparation process of the cores is crucial to their performance, especially in the sintering and annealing stages. Annealing is a key step in optimizing the performance of the cores. It heats the material to improve its lattice structure, reduce internal stress and enhance magnetic properties.

[0003] During traditional sintering and annealing processes, temperature control systems often rely on a preset, uniform temperature curve, failing to account for temperature differences at different locations within the furnace and their impact on the core sintering process. Furthermore, due to factors such as uneven airflow distribution within the furnace and an irrational layout of heating elements, large temperature fluctuations can lead to variations in the magnetic properties of the cores, making it difficult to ensure consistent performance across batches, impacting the overall quality of the cores. Summary of the Invention

[0004] The purpose of the present invention is to provide a temperature control optimization method, system and equipment for semiconductor magnetic core preparation, which is used to solve the technical problems of insufficient annealing control accuracy and delayed temperature control response in the traditional semiconductor magnetic core preparation process, including: In a first aspect, the present invention provides a temperature control optimization method for the preparation of semiconductor magnetic cores, comprising: configuring a temperature monitoring array according to the temperature zoning results of a high-temperature furnace; during the sintering process of the semiconductor magnetic core, performing regular monitoring through the temperature monitoring array to obtain a plurality of temperature data sequences; predicting and obtaining a plurality of magnetic core sintering information based on the plurality of temperature data sequences; configuring an annealing temperature control scheme according to the plurality of magnetic core sintering information, and performing annealing temperature zoning control based on the temperature zoning results.

[0005] Preferably, the temperature control optimization method for the preparation of semiconductor magnetic cores also includes: obtaining the heating element layout, air flow distribution in the furnace, furnace body material and furnace body size of the high-temperature furnace; obtaining product attribute information of the semiconductor magnetic core to be sintered, and product distribution information in the furnace; dividing the temperature area in the furnace according to the heating element layout, air flow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information, and obtaining temperature zoning results, wherein the temperature zoning results include multiple furnace areas; and arranging temperature sensors in each furnace area to construct a temperature monitoring array.

[0006] Preferably, the temperature control optimization method for the preparation of semiconductor magnetic cores also includes: retrieving historical product sintering logs based on the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints to obtain multiple temperature distribution arrays; dividing the multiple temperature distribution arrays according to the historical temperature error mean of the temperature sensor, clustering the temperature coverage areas with a temperature difference less than the historical temperature error mean into the same area, and obtaining multiple area division results; selecting the mode of the multiple area division results as the temperature partition result.

[0007] Preferably, the temperature control optimization method for semiconductor magnetic core preparation also includes: randomly selecting a first temperature data sequence from the several temperature data sequences; performing temperature fluctuation and temperature deviation analysis on the first temperature data sequence to obtain a first fluctuation coefficient and a first deviation coefficient; pre-training a magnetic core sintering predictor; using the magnetic core sintering predictor to predict first magnetic core sintering information based on the first temperature data sequence, the first fluctuation coefficient and the first deviation coefficient, and adding the first magnetic core sintering information to the several magnetic core sintering information.

[0008] Preferably, the temperature control optimization method for the preparation of semiconductor magnetic cores also includes: calculating the temperature mean and temperature standard deviation of the first temperature data sequence, and setting the ratio of the temperature standard deviation to the temperature mean as a first fluctuation coefficient; taking the standard sintering temperature sequence of the semiconductor magnetic core as a benchmark, performing temperature deviation analysis on the same node according to the first temperature data sequence, obtaining multiple deviation values, and calculating the average to obtain the first deviation coefficient.

[0009] Preferably, the temperature control optimization method for the preparation of semiconductor magnetic cores also includes: using the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints, retrieving historical product sintering logs, collecting sample temperature sequence sets, sample fluctuation coefficient sets, sample deviation coefficient sets and sample sintering information sets, wherein the sintering information includes grain size and particle density; using the sample temperature sequence, sample fluctuation coefficient and sample deviation coefficient as input, and using the sample sintering information as supervision, using the sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set to train a feedforward neural network until the model converges, thereby obtaining the magnetic core sintering predictor.

[0010] Preferably, the temperature control optimization method for semiconductor magnetic core preparation also includes: matching in the initial annealing temperature database according to the plurality of magnetic core sintering information to obtain a plurality of initial annealing temperatures; obtaining a plurality of deviation coefficients corresponding to the plurality of magnetic core sintering information, and setting a plurality of feedback control frequencies according to the plurality of deviation coefficients; generating an annealing temperature control scheme according to the plurality of initial annealing temperatures and the plurality of feedback control frequencies.

[0011] Preferably, the temperature control optimization method for the preparation of semiconductor magnetic cores also includes: setting the deviation coefficient greater than a predetermined deviation threshold as a risk deviation coefficient, counting the proportion of the risk deviation coefficient in the several deviation coefficients to obtain the deviation ratio, and constructing a risk deviation coefficient distribution; performing risk deviation discrete analysis based on the risk deviation coefficient distribution to obtain the risk deviation concentration; if the deviation ratio is greater than the predetermined ratio threshold and / or the risk deviation concentration is greater than the predetermined concentration, setting the predetermined maximum control frequency as the feedback control frequency to obtain several feedback control frequencies; if the deviation ratio is less than the predetermined ratio threshold and the risk deviation concentration is less than the predetermined concentration, setting several feedback control frequencies according to the several deviation coefficients, wherein the feedback control frequency is the ratio of the deviation coefficient to the historical maximum deviation coefficient multiplied by the predetermined maximum control frequency.

[0012] In the second aspect, the present invention also provides a temperature control optimization system for the preparation of semiconductor magnetic cores, which is used to execute a temperature control optimization method for the preparation of semiconductor magnetic cores as described in the first aspect, including: a temperature monitoring array configuration module, which is used to configure the temperature monitoring array according to the temperature zoning results of the high-temperature furnace; a temperature data sequence acquisition module, which is used to perform regular monitoring through the temperature monitoring array during the sintering process of the semiconductor magnetic core to obtain a number of temperature data sequences; a magnetic core sintering information prediction module, which is used to obtain a number of magnetic core sintering information based on the prediction of the several temperature data sequences; an annealing temperature zoning control module, which is used to configure an annealing temperature control scheme according to the several magnetic core sintering information, and perform annealing temperature zoning control based on the temperature zoning results.

[0013] In a third aspect, the present invention further provides an electronic device, comprising: At least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so as to enable the at least one processor to perform the steps of any one of the methods described in the first aspect above.

[0014] The embodiments of the present invention include the following advantages: By configuring a temperature monitoring array based on the temperature zoning results of a high-temperature furnace; during the sintering process of a semiconductor magnetic core, regular monitoring is performed through the temperature monitoring array to obtain a number of temperature data sequences; then, a number of magnetic core sintering information is predicted based on the several temperature data sequences; then, an annealing temperature control scheme is configured based on the several magnetic core sintering information; and finally, annealing temperature zoning control is performed based on the temperature zoning results. That is, through precise temperature monitoring and zoning control, the temperature uniformity and stability of the annealing process can be improved, thereby optimizing the performance and consistency of the magnetic core, reducing the defect rate, and improving production efficiency and energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the steps of a temperature control optimization method for preparing a semiconductor magnetic core according to the present invention; Figure 2 This is a schematic structural diagram of a temperature control optimization system for semiconductor magnetic core preparation according to the present invention; Figure 3 This is a schematic structural diagram of the electronic device provided by the present invention.

[0016] Description of reference numerals: Temperature monitoring array configuration module 11 , temperature data sequence acquisition module 12 , magnetic core sintering information prediction module 13 , annealing temperature partition control module 14 , electronic device 500 , memory 510 , processor 520 , first computer program 511 . DETAILED DESCRIPTION

[0017] This invention addresses the technical issues of insufficient annealing control accuracy and delayed temperature control response during conventional semiconductor core fabrication by providing a temperature control optimization method, system, and equipment for semiconductor core fabrication. Through precise temperature monitoring and zoned control, the temperature uniformity and stability of the annealing process can be improved, thereby optimizing the performance and consistency of the core, reducing defect rates, and increasing production efficiency and energy efficiency.

[0018] Below, the technical solutions of the present invention will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments described herein. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should also be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the accompanying drawings.

[0019] For example, see the attached Figure 1The present invention provides a temperature control optimization method for semiconductor magnetic core preparation, which is applied to a temperature control optimization system for semiconductor magnetic core preparation, and specifically includes the following steps: S10: configuring a temperature monitoring array according to the temperature zoning result of the high temperature furnace.

[0020] Furthermore, step S10 of the present invention further includes: S11: Obtain the heating element layout, airflow distribution in the high-temperature furnace, furnace body material and furnace body size; S12: Obtain product attribute information of the semiconductor core to be sintered, and product distribution information in the furnace.

[0021] Specifically, the layout of the high-temperature furnace's heating elements, airflow distribution within the furnace, and the furnace material and size are determined. Different types of high-temperature furnaces have different heating element locations and power distributions. For example, resistance wire heating, induction heating, or infrared heating each have different temperature gradient distributions. Therefore, the specific layout of the heating elements needs to be determined to assess the heating intensity of each area. The flow pattern of airflow within the furnace (e.g., natural convection, forced convection) directly affects temperature uniformity. Reasonable airflow design helps reduce temperature gradients and improve temperature control accuracy. The material of the furnace wall (e.g., ceramic fiber, refractory brick, stainless steel) affects heat reflection, heat conduction, and heat loss within the furnace. The thermal conductivity of different materials determines the stability of the temperature distribution. The size of the furnace directly determines the complexity of the temperature distribution. Larger furnaces may have more significant temperature gradients, while smaller furnaces are relatively easier to control temperature uniformity. These data will serve as the basis for subsequent temperature zoning to ensure that the temperature difference in each area is within a controllable range to optimize the sintering and annealing processes.

[0022] Secondly, information on the product attributes of the semiconductor cores to be sintered, as well as the distribution of the cores within the furnace, is obtained. This information includes material composition, geometry, and core density. Magnetic cores of different materials (such as ferrite, amorphous alloy, and nanocrystalline) have varying temperature tolerances, resulting in distinct sintering and annealing temperature profiles. The size and shape of the cores affect heat conduction and shrinkage during sintering, necessitating temperature optimization for cores of varying specifications. Cores of varying densities experience varying degrees of thermal expansion and contraction during sintering, requiring temperature control adjustments to account for this factor. Product distribution information includes product placement and batch size. The arrangement of the cores within the furnace (such as stacking, height, and spacing) influences local temperature gradients; dense stacking can lead to localized overheating or cold spots. The number of cores loaded within the furnace determines the overall heat capacity, impacting the temperature stability within the furnace. This data is used to optimize the layout of the temperature monitoring array and, combined with the temperature monitoring data, dynamically adjust the sintering and annealing temperature control schemes to improve the consistency and stability of the core performance.

[0023] S13: Divide the temperature zones in the furnace according to the heating element layout, the airflow distribution in the furnace, the furnace body material, the furnace body size, the product attribute information and the product distribution information, and obtain temperature zoning results, wherein the temperature zoning results include multiple furnace zones.

[0024] Furthermore, step S13 of the present invention further includes: S131: With the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints, retrieve the historical product sintering log and obtain multiple temperature distribution arrays; S132: Divide the multiple temperature distribution arrays according to the historical temperature error mean of the temperature sensor, cluster the temperature coverage areas with a temperature difference less than the historical temperature error mean into the same area, and obtain multiple area division results; S133: Select the mode of the multiple area division results and set it as the temperature partition result.

[0025] Specifically, first, obtain the layout of heating elements, airflow distribution in the furnace, furnace material and furnace size. This information determines the temperature gradient in the furnace and the distribution of local heat sources. For example, the position of the heating elements and the flow pattern of the airflow will affect the temperature distribution in different areas. Therefore, these data are needed as constraints to screen and analyze historical sintering data; and obtain product attribute information and product distribution information. Factors such as the material, geometric size, density of the magnetic core, and the distribution method in the furnace (such as uniform distribution or concentrated stacking) will affect the temperature distribution. For example, some products are stacked densely, which may cause local temperatures to be too high.

[0026] Next, using the heating element layout, furnace airflow distribution, furnace material, furnace size, product attribute information, and product distribution information as constraints, historical product sintering logs are retrieved. By retrieving historical sintering data (temperature distribution, time, furnace environment, and other information), multiple different temperature distribution arrays can be obtained; these temperature distribution arrays reflect the temperature changes in the furnace during past production processes and can provide valuable reference for current production.

[0027] Then, the historical temperature error mean of the temperature sensor is obtained, and the average value of the sensor temperature measurement error calculated through historical data is the maximum tolerable error of temperature monitoring; further, for each temperature distribution array, the area is divided according to the temperature difference. If the temperature difference of some areas is less than the historical temperature error mean, these areas are aggregated together to form a larger area to ensure that the temperature difference in each area is within the allowable error range, reduce unnecessary area divisions, and ensure that the temperature monitoring system works efficiently within a reasonable range; obtain multiple area division results. Finally, the mode of the multiple area division results is selected as the temperature partition result, that is, the most frequently occurring area division result is used as the final temperature partition scheme. Multiple possible temperature distribution arrays are inferred through historical sintering data, and then the areas with temperature differences less than the mean are merged into the same area according to the sensor error mean. Finally, the temperature partition result that best suits the current production environment is selected through mode statistics to provide data support for precise temperature control and subsequent temperature monitoring.

[0028] S14: Temperature sensors are arranged in each area of the furnace to construct a temperature monitoring array.

[0029] Specifically, in the previous steps, based on historical temperature distribution data, mean error, and temperature zoning analysis, the furnace is divided into multiple temperature zones (such as high temperature, low temperature, and uniform temperature). Each zone has distinct temperature characteristics and trends. Therefore, deploying temperature sensors in each zone is crucial for ensuring accurate temperature monitoring. Next, the appropriate sensor density is selected based on the temperature fluctuation amplitude and temperature differential within each zone. In areas with large temperature fluctuations, more sensors are deployed to ensure accurate capture of temperature fluctuations. In areas with smaller temperature differentials or relatively stable temperatures, sensors can be deployed more sparsely. Finally, the temperature data acquired by each sensor is aggregated to form a temperature monitoring array. Each array represents the temperature distribution of a zone, and the data from each sensor in the array forms an integrated temperature monitoring network. Deploying temperature sensors in each furnace zone and constructing a temperature monitoring array provides high-precision data support for the entire temperature control process. By acquiring real-time temperature information from each zone, control parameters for the heating and annealing processes can be dynamically adjusted to ensure that the temperature of each zone remains within the optimal range, thereby optimizing the performance and consistency of the semiconductor core.

[0030] S20: During the sintering process of the semiconductor core, regular monitoring is performed through the temperature monitoring array to obtain a plurality of temperature data sequences.

[0031] Specifically, during the sintering process of the semiconductor core, regular monitoring is performed through the temperature monitoring array, that is, the temperature of each area is sampled at a preset time interval (such as every second or every minute). The selection of the sampling frequency needs to balance data accuracy and computational burden to ensure that the system can respond to temperature changes in real time and provide sufficient data for subsequent analysis. Since multiple temperature sensors may be deployed in different areas, multiple sensors in the same area will provide multiple temperature data at the same time point. In order to reduce the impact of the measurement error of a single sensor on temperature control, the temperature value of each area uses the average of the temperature measurement values of multiple sensors in the area to represent the overall temperature of the area. As time goes by, the temperature average of each area will be continuously recorded, thereby forming a temperature data sequence, and obtaining several temperature data sequences.

[0032] S30: Predict and obtain a plurality of magnetic core sintering information according to the plurality of temperature data sequences.

[0033] Furthermore, step S30 of the present invention further includes: S31: Randomly select a first temperature data sequence from the plurality of temperature data sequences.

[0034] Specifically, any one temperature data sequence is randomly selected from the plurality of temperature data sequences and set as the first temperature data sequence.

[0035] S32: Performing temperature fluctuation and temperature deviation analysis on the first temperature data sequence to obtain a first fluctuation coefficient and a first deviation coefficient.

[0036] Furthermore, step S32 of the present invention further includes: S321: Calculate the temperature mean and temperature standard deviation of the first temperature data sequence, and set the ratio of the temperature standard deviation to the temperature mean as the first fluctuation coefficient; S322: Based on the standard sintering temperature sequence of the semiconductor magnetic core, perform temperature deviation analysis on the same node according to the first temperature data sequence, obtain multiple deviation values, and calculate the average to obtain the first deviation coefficient.

[0037] Specifically, first, the temperature mean and temperature standard deviation of multiple first temperature data in the first temperature data sequence are calculated, and the ratio of the temperature standard deviation to the temperature mean is set as the first fluctuation coefficient. The first fluctuation coefficient is used to measure the stability of the temperature in the area. If the first fluctuation coefficient is too high, it means that the temperature fluctuation in the area is large and the temperature control strategy needs to be adjusted; if the first fluctuation coefficient is too low, it means that the temperature in the area is relatively stable, which is conducive to the stability of the sintering process.

[0038] Next, a standard sintering temperature sequence of the semiconductor magnetic core is obtained, that is, the target temperature at each time point under ideal circumstances; then, based on the standard sintering temperature sequence, a temperature deviation analysis of the same node is performed according to the first temperature data sequence, and the ratio of the temperature difference value at the same node to the standard sintering temperature is set as the deviation value, and multiple deviation values are obtained. The average of the multiple deviation values is calculated to obtain a first deviation coefficient, which measures the average deviation degree of the overall temperature data sequence relative to the standard temperature.

[0039] S33: Pre-trained magnetic core sintering predictor.

[0040] Furthermore, step S33 of the present invention further includes: S331: With the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints, retrieve historical product sintering logs, collect sample temperature sequence sets, sample fluctuation coefficient sets, sample deviation coefficient sets and sample sintering information sets, wherein the sintering information includes grain size and particle density; S332: With the sample temperature sequence, sample fluctuation coefficient and sample deviation coefficient as input, and the sample sintering information as supervision, use the sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set to train a feedforward neural network until the model converges, thereby obtaining the magnetic core sintering predictor.

[0041] Specifically, based on the constraints of the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information, the historical product sintering log is retrieved, that is, the relevant sintering log is retrieved from the historical database. The sintering log contains detailed records of temperature monitoring, fluctuations, deviations and sintering results in the previous sintering process. The sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set are used. Among them, the temperature data sequence of each historical sample records the temperature changes of each area in the furnace during the sintering process; the temperature fluctuation coefficient of each sample is used to describe the stability and volatility of the temperature; the temperature deviation coefficient of each sample is used to measure the deviation between the actual temperature and the standard temperature; the sintering results corresponding to each sample include grain size and particle density. These two indicators are important parameters for evaluating the sintering quality of the magnetic core.

[0042] Next, the sample temperature sequence, sample fluctuation coefficient and sample deviation coefficient are used as input, and the sample sintering information is used as supervision. The sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set are used to train a feedforward neural network, wherein the input layer of the neural network will receive the sample temperature sequence, fluctuation coefficient and deviation coefficient, and these input data are passed to the next layer of the network through feature extraction or encoding; the feedforward neural network usually contains multiple hidden layers, and the neurons in each layer will be weighted and summed, and then the output will be generated through the activation function; the output layer will make predictions based on the sintering results (grain size and particle density) of the training data. Neural networks are trained through forward propagation and backpropagation. During each forward propagation, the input data passes through the various layers of the network to obtain a predicted output. The error is then evaluated by calculating the loss function. The loss function is the most critical part of the training process. It measures the difference between the model's predicted value and the actual value, for example, the mean squared error loss function. The backpropagation algorithm updates the weights and biases in the network based on the gradient information of the loss function. Through multiple iterations, the model gradually converges and the loss function tends to a minimum value. When the value of the loss function does not change much in consecutive rounds of iterations, it indicates that the model has converged and training can be stopped. At this time, the trained magnetic core sintering predictor is obtained.

[0043] By adjusting the weights of the neural network through multiple iterations, it is possible to predict the changes in grain size and particle density during the sintering process based on inputs such as temperature data, fluctuation coefficient and deviation coefficient; this training method provides an accurate prediction tool for the sintering process of semiconductor magnetic cores, which can effectively optimize the sintering process and improve product quality.

[0044] S34: Utilizing the magnetic core sintering predictor, predicting first magnetic core sintering information according to the first temperature data sequence, the first fluctuation coefficient, and the first deviation coefficient, and adding the first magnetic core sintering information to the plurality of magnetic core sintering information.

[0045] Specifically, the first temperature data sequence, the first fluctuation coefficient and the first deviation coefficient are input into the magnetic core sintering predictor for prediction, and the first magnetic core sintering information is output and added to the plurality of magnetic core sintering information.

[0046] S40: configuring an annealing temperature control scheme according to the plurality of magnetic core sintering information, and performing annealing temperature zoning control based on the temperature zoning result.

[0047] Furthermore, step S40 of the present invention further includes: S41: performing matching in an initial annealing temperature database according to the plurality of magnetic core sintering information to obtain a plurality of initial annealing temperatures.

[0048] Specifically, an initial annealing temperature database is constructed, and the initial annealing temperature database contains records of multiple annealing temperatures, which are pre-set based on different sintering information (such as grain size, particle density, etc.). Each record in the database will contain sintering information and corresponding annealing temperatures. These temperatures are optimized based on experience or experimental data, which can ensure that the performance of the magnetic core is further improved during the annealing process. Then, the sintering information of the several magnetic cores is input into the initial annealing temperature database for matching. In order to ensure that it matches the information in the database, the sintering information (such as grain size and particle density) can be standardized. This can be done by normalizing the grain size and particle density so that its range adapts to the information format in the database; using the nearest neighbor algorithm or Euclidean distance calculation, the sintering information of the magnetic core is compared with the known sintering information in the database, the most similar sintering information is selected, and several initial annealing temperatures are output, wherein the initial annealing temperature database is shown in Table 1: Table 1: Initial annealing temperature database ; ; By matching with historical data, a suitable annealing temperature can be quickly selected for the sintering process; through this matching process, precise control of the annealing temperature can be achieved, thereby further improving the performance and consistency of the magnetic core and reducing deviations and defects in the production process.

[0049] S42: Acquire a plurality of deviation coefficients corresponding to a plurality of magnetic core sintering information, and set a plurality of feedback control frequencies according to the plurality of deviation coefficients.

[0050] Furthermore, step S42 of the present invention further includes: S421: Set the deviation coefficient greater than the predetermined deviation threshold as the risk deviation coefficient, count the proportion of the risk deviation coefficient in the several deviation coefficients, obtain the deviation ratio, and construct the risk deviation coefficient distribution; S422: Perform risk deviation discrete analysis based on the risk deviation coefficient distribution to obtain the risk deviation concentration; S423: If the deviation ratio is greater than the predetermined ratio threshold and / or the risk deviation concentration is greater than the predetermined concentration, set the predetermined maximum control frequency as the feedback control frequency to obtain several feedback control frequencies; S424: If the deviation ratio is less than the predetermined ratio threshold and the risk deviation concentration is less than the predetermined concentration, set several feedback control frequencies according to the several deviation coefficients, wherein the feedback control frequency is the ratio of the deviation coefficient to the historical maximum deviation coefficient multiplied by the predetermined maximum control frequency.

[0051] Specifically, during the sintering process, temperature deviations can affect the final performance of the magnetic core. By calculating the deviation coefficient of each temperature data sequence (such as the deviation from the standard sintering temperature), the accuracy of temperature control can be determined. If a deviation coefficient exceeds a predetermined deviation threshold (for example, a deviation outside a set range), it is marked as a risk deviation coefficient, indicating that this area or temperature point has a high risk. Next, all deviation coefficients are statistically analyzed to calculate the proportion of risk deviation coefficients exceeding the predetermined deviation threshold to all deviation coefficients. This proportion reflects the stability of temperature control throughout the sintering process. By analyzing all risk deviation coefficients, a risk deviation coefficient distribution (including regional location information) is generated to further understand which areas or time periods experience significant fluctuations in temperature control.

[0052] Next, based on the distribution of risk deviation coefficients, a discrete analysis is performed to observe whether the distribution of the deviation coefficients is uniform. If most of the risk deviation coefficients are concentrated in certain specific areas or time periods, it means that the temperature control problem is relatively concentrated; the risk deviation concentration indicates whether the areas with large temperature fluctuations are concentrated in a few locations or time periods. The higher the concentration, the more serious the temperature control problems in these areas may be, and they need to be focused on. Configure a predetermined concentration. The predetermined concentration is a set concentration value. When the concentration exceeds this value, it means that the areas where the temperature deviation is more concentrated need to adjust the temperature more frequently. If the deviation ratio is greater than the predetermined ratio threshold and / or the risk deviation concentration is greater than the predetermined concentration, it means that the system has a more serious temperature control problem. At this time, the feedback control frequency should be increased, and the maximum control frequency should be used for adjustment to correct the temperature control deviation more frequently.

[0053] If the deviation ratio is less than the predetermined ratio threshold and the risk deviation concentration is less than the predetermined concentration, the system temperature control is relatively stable. In this case, the feedback control frequency should be dynamically adjusted based on the deviation coefficient of each area. The current deviation coefficient is compared with the maximum deviation coefficient in the historical record to calculate the adjustment ratio, and the control frequency is set based on this. The feedback control frequency is the ratio of the deviation coefficient to the historical maximum deviation coefficient multiplied by the predetermined maximum control frequency, resulting in a number of feedback control frequencies. In this way, the temperature control system can be moderately adjusted when the deviation is small, avoiding unnecessary over-adjustment while ensuring that the temperature control remains flexible within a reasonable range.

[0054] By statistically analyzing and identifying potential risk areas in temperature control, and adjusting the feedback control frequency accordingly, the accuracy of annealing temperature control during the sintering process of semiconductor cores can be ensured. This method can effectively cope with different temperature control deviations and reduce quality problems caused by improper temperature control during the production process.

[0055] S43: generating an annealing temperature control scheme according to the plurality of initial annealing temperatures and the plurality of feedback control frequencies.

[0056] Specifically, a complete annealing temperature control plan is generated by combining the initial annealing temperature and the feedback control frequency. This plan includes the temperature setting for each stage and the corresponding temperature control frequency. Finally, annealing temperature is controlled based on the temperature partitioning results according to the annealing temperature control plan.

[0057] In summary, the temperature control optimization method for semiconductor magnetic core preparation provided by the present invention has the following technical effects: By configuring a temperature monitoring array based on the temperature zoning results of a high-temperature furnace; during the sintering process of a semiconductor magnetic core, regular monitoring is performed through the temperature monitoring array to obtain a number of temperature data sequences; then, a number of magnetic core sintering information is predicted based on the several temperature data sequences; then, an annealing temperature control scheme is configured based on the several magnetic core sintering information; and finally, annealing temperature zoning control is performed based on the temperature zoning results. That is, through precise temperature monitoring and zoning control, the temperature uniformity and stability of the annealing process can be improved, thereby optimizing the performance and consistency of the magnetic core, reducing the defect rate, and improving production efficiency and energy efficiency.

[0058] In the second embodiment, based on the same inventive concept as the temperature control optimization method for preparing a semiconductor magnetic core in the above embodiment, the present invention also provides a temperature control optimization system for preparing a semiconductor magnetic core, please refer to the attached Figure 2 , including: a temperature monitoring array configuration module 11, used to configure the temperature monitoring array according to the temperature zoning result of the high-temperature furnace; a temperature data sequence acquisition module 12, used to perform regular monitoring through the temperature monitoring array during the semiconductor core sintering process to obtain a number of temperature data sequences; a core sintering information prediction module 13, used to predict and obtain a number of core sintering information based on the several temperature data sequences; an annealing temperature zoning control module 14, used to configure the annealing temperature control scheme according to the several core sintering information, and perform annealing temperature zoning control based on the temperature zoning results.

[0059] Furthermore, the temperature control optimization system for the preparation of semiconductor magnetic cores is also used to: obtain the heating element layout, air flow distribution in the furnace, furnace body material and furnace body size of the high-temperature furnace; obtain product attribute information of the semiconductor magnetic core to be sintered, and product distribution information in the furnace; divide the temperature area in the furnace according to the heating element layout, air flow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information, and obtain temperature zoning results, wherein the temperature zoning results include multiple furnace areas; and arrange temperature sensors in each furnace area to construct a temperature monitoring array.

[0060] Furthermore, the temperature control optimization system for semiconductor core preparation is also used to: retrieve historical product sintering logs based on the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints to obtain multiple temperature distribution arrays; divide the multiple temperature distribution arrays according to the historical temperature error mean of the temperature sensor, cluster the temperature coverage areas with temperature differences less than the historical temperature error mean into the same area, and obtain multiple area division results; select the mode of the multiple area division results as the temperature partition result.

[0061] Furthermore, the temperature control optimization system for semiconductor magnetic core preparation is also used to: randomly select a first temperature data sequence from the several temperature data sequences; perform temperature fluctuation and temperature deviation analysis on the first temperature data sequence to obtain a first fluctuation coefficient and a first deviation coefficient; pre-train a magnetic core sintering predictor; and use the magnetic core sintering predictor to predict first magnetic core sintering information based on the first temperature data sequence, the first fluctuation coefficient and the first deviation coefficient, and add the first magnetic core sintering information to the several magnetic core sintering information.

[0062] Furthermore, the temperature control optimization system for semiconductor magnetic core preparation is also used to: calculate the temperature mean and temperature standard deviation of the first temperature data sequence, and set the ratio of the temperature standard deviation to the temperature mean as a first fluctuation coefficient; based on the standard sintering temperature sequence of the semiconductor magnetic core, perform temperature deviation analysis on the same node according to the first temperature data sequence, obtain multiple deviation values, and calculate the average to obtain the first deviation coefficient.

[0063] Furthermore, the temperature control optimization system for semiconductor magnetic core preparation is also used to: retrieve historical product sintering logs based on the heating element layout, airflow distribution in the furnace, furnace body material, furnace body size, product attribute information and product distribution information as constraints, collect sample temperature sequence sets, sample fluctuation coefficient sets, sample deviation coefficient sets and sample sintering information sets, wherein the sintering information includes grain size and particle density; use the sample temperature sequence, sample fluctuation coefficient and sample deviation coefficient as input, and use the sample sintering information as supervision, and use the sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set to train a feedforward neural network until the model converges to obtain the magnetic core sintering predictor.

[0064] Furthermore, the temperature control optimization system for semiconductor magnetic core preparation is also used to: match the initial annealing temperature database based on the multiple magnetic core sintering information to obtain multiple initial annealing temperatures; obtain multiple deviation coefficients corresponding to the multiple magnetic core sintering information, and set multiple feedback control frequencies based on the multiple deviation coefficients; generate an annealing temperature control plan based on the multiple initial annealing temperatures and the multiple feedback control frequencies.

[0065] Furthermore, the temperature control optimization system for the preparation of semiconductor magnetic cores is also used to: set the deviation coefficient greater than a predetermined deviation threshold as a risk deviation coefficient, count the proportion of the risk deviation coefficient in the several deviation coefficients, obtain the deviation ratio, and construct a risk deviation coefficient distribution; perform risk deviation discrete analysis based on the risk deviation coefficient distribution to obtain the risk deviation concentration; if the deviation ratio is greater than the predetermined ratio threshold and / or the risk deviation concentration is greater than the predetermined concentration, set the predetermined maximum control frequency as the feedback control frequency to obtain several feedback control frequencies; if the deviation ratio is less than the predetermined ratio threshold and the risk deviation concentration is less than the predetermined concentration, set several feedback control frequencies according to the several deviation coefficients, wherein the feedback control frequency is the ratio of the deviation coefficient to the historical maximum deviation coefficient multiplied by the predetermined maximum control frequency.

[0066] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The temperature control optimization method and specific examples for preparing semiconductor magnetic cores in the aforementioned embodiment 1 are also applicable to a temperature control optimization system for preparing semiconductor magnetic cores in this embodiment. Through the aforementioned detailed description of a temperature control optimization method for preparing semiconductor magnetic cores, those skilled in the art can clearly understand the temperature control optimization system for preparing semiconductor magnetic cores in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0067] For example three, please refer to Figure 3 , Figure 3 Schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a first computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the first computer program 511, the following steps are implemented: configuring a temperature monitoring array according to the temperature zoning result of the high-temperature furnace; performing regular monitoring through the temperature monitoring array during the sintering process of the semiconductor magnetic core to obtain a plurality of temperature data sequences; obtaining a plurality of magnetic core sintering information based on the prediction of the plurality of temperature data sequences; configuring an annealing temperature control scheme according to the plurality of magnetic core sintering information, and performing annealing temperature zoning control based on the temperature zoning result.

[0068] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0069] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is intended to include these modifications and variations.

Claims

1. A temperature control optimization method for semiconductor magnetic core preparation, characterized in that: Methods include: Configure the temperature monitoring array according to the temperature zoning results of the high-temperature furnace; During the sintering process of the semiconductor core, the temperature monitoring array is used to perform regular monitoring to obtain a plurality of temperature data sequences; Predicting and obtaining a plurality of magnetic core sintering information according to the plurality of temperature data sequences; An annealing temperature control scheme is configured according to the plurality of magnetic core sintering information, and annealing temperature zoning control is performed based on the temperature zoning results.

2. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 1, characterized in that: Configure the temperature monitoring array based on the temperature zoning results of the high-temperature furnace, including: Obtain the heating element layout, air flow distribution, furnace material and furnace size of the high-temperature furnace; Obtain product attribute information of the semiconductor core to be sintered, as well as product distribution information in the furnace; Dividing the furnace into temperature zones according to the heating element layout, the airflow distribution in the furnace, the furnace material, the furnace size, the product attribute information, and the product distribution information to obtain a temperature zone result, wherein the temperature zone result includes a plurality of furnace zones; Temperature sensors are placed in each area of the furnace to construct a temperature monitoring array.

3. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 2, characterized in that: The temperature zones in the furnace are divided according to the heating element layout, airflow distribution in the furnace, furnace material, furnace size, product attribute information, and product distribution information, including: Using the heating element layout, furnace airflow distribution, furnace material, furnace size, product attribute information, and product distribution information as constraints, retrieve historical product sintering logs to obtain multiple temperature distribution arrays; Dividing the multiple temperature distribution arrays according to historical temperature error mean values of the temperature sensors, clustering temperature coverage areas with temperature differences less than the historical temperature error mean values into the same area, and obtaining multiple area division results; The mode of the plurality of region division results is selected as the temperature partition result.

4. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 2, characterized in that: Predicting and obtaining a plurality of magnetic core sintering information according to the plurality of temperature data sequences includes: randomly selecting a first temperature data sequence from the plurality of temperature data sequences; performing temperature fluctuation and temperature deviation analysis on the first temperature data sequence to obtain a first fluctuation coefficient and a first deviation coefficient; Pre-trained core sintering predictor; The magnetic core sintering predictor is used to predict first magnetic core sintering information based on the first temperature data sequence, the first fluctuation coefficient and the first deviation coefficient, and the first magnetic core sintering information is added to the plurality of magnetic core sintering information.

5. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 4, characterized in that: Performing temperature fluctuation and temperature deviation analysis on the first temperature data sequence includes: Calculating a temperature mean and a temperature standard deviation of the first temperature data sequence, and setting a ratio of the temperature standard deviation to the temperature mean as a first fluctuation coefficient; Based on the standard sintering temperature sequence of the semiconductor magnetic core, a temperature deviation analysis of the same node is performed according to the first temperature data sequence to obtain multiple deviation values, and the first deviation coefficient is obtained by averaging.

6. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 4, characterized in that: Pre-trained core sintering predictor, including: Using the heating element layout, airflow distribution in the furnace, furnace material, furnace size, product attribute information, and product distribution information as constraints, historical product sintering logs are retrieved to collect a sample temperature sequence set, a sample fluctuation coefficient set, a sample deviation coefficient set, and a sample sintering information set, wherein the sintering information includes grain size and particle density; The sample temperature sequence, sample fluctuation coefficient and sample deviation coefficient are used as inputs and the sample sintering information is used as supervision. The feedforward neural network is trained using the sample temperature sequence set, sample fluctuation coefficient set, sample deviation coefficient set and sample sintering information set until the model converges, thereby obtaining the magnetic core sintering predictor.

7. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 5, characterized in that: Configuring an annealing temperature control scheme according to the plurality of magnetic core sintering information includes: According to the plurality of magnetic core sintering information, matching is performed in an initial annealing temperature database to obtain a plurality of initial annealing temperatures; Acquiring a plurality of deviation coefficients corresponding to a plurality of magnetic core sintering information, and setting a plurality of feedback control frequencies according to the plurality of deviation coefficients; An annealing temperature control scheme is generated according to the several initial annealing temperatures and the several feedback control frequencies.

8. The temperature control optimization method for preparing a semiconductor magnetic core according to claim 7, characterized in that: Setting a plurality of feedback control frequencies according to the plurality of deviation coefficients includes: The deviation coefficient greater than the predetermined deviation threshold is set as the risk deviation coefficient, the proportion of the risk deviation coefficient in the plurality of deviation coefficients is counted to obtain the deviation ratio, and the risk deviation coefficient distribution is constructed; Performing risk deviation dispersion analysis based on the risk deviation coefficient distribution to obtain risk deviation concentration; If the deviation ratio is greater than a predetermined ratio threshold and / or the risk deviation concentration is greater than a predetermined concentration, the predetermined maximum control frequency is set as the feedback control frequency to obtain a plurality of feedback control frequencies; If the deviation ratio is less than the predetermined ratio threshold and the risk deviation concentration is less than the predetermined concentration, several feedback control frequencies are set according to several deviation coefficients, wherein the feedback control frequency is the ratio of the deviation coefficient to the historical maximum deviation coefficient multiplied by the predetermined maximum control frequency.

9. A temperature control optimization system for semiconductor magnetic core preparation, characterized in that: The steps for implementing the temperature control optimization method for preparing a semiconductor magnetic core according to any one of claims 1 to 8 include: A temperature monitoring array configuration module is used to configure the temperature monitoring array according to the temperature zoning result of the high-temperature furnace; a temperature data sequence acquisition module, configured to periodically monitor the temperature during the sintering process of the semiconductor core through the temperature monitoring array to acquire a plurality of temperature data sequences; A magnetic core sintering information prediction module, configured to predict and obtain a plurality of magnetic core sintering information according to the plurality of temperature data sequences; The annealing temperature zoning control module is used to configure an annealing temperature control scheme according to the plurality of magnetic core sintering information, and perform annealing temperature zoning control based on the temperature zoning results.

10. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the steps of the temperature control optimization method for preparing a semiconductor magnetic core as described in any one of claims 1 to 8.

Images