Oxygen content prediction method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide an oxygen content prediction method, device, equipment and medium. The method comprises: in the current single crystal pulling process, obtaining characteristic data before the head stage of the isochronal growth, wherein the characteristic data comprises operation data related to the oxygen content of the head stage of the isochronal growth, inputting the characteristic data into an oxygen content prediction model, wherein the oxygen content prediction model is trained by characteristic data samples before the head stage of the isochronal growth and sample oxygen contents of the corresponding labeled head stage of the isochronal growth, generating a predicted oxygen content of the head stage of the isochronal growth from the oxygen content prediction model according to the characteristic data, so that the oxygen content of the head stage of the isochronal growth can be predicted in advance, thereby the oxygen content can be controlled in advance, and the crystal quality is improved.

Description

Technical Field

[0001] This invention relates to the field of crystal preparation technology, and in particular to an oxygen content prediction method, an oxygen content prediction device, an electronic device, and a storage medium. Background Technology

[0002] The main process for preparing monocrystalline silicon materials is the Czochralski process (CZ), which uses the Czochralski method to refine polycrystalline silicon raw materials into monocrystalline silicon. The process of generating rod-shaped monocrystalline silicon crystals during the Czochralski single crystal growth process includes steps such as charging, melting, feeding, temperature control, crystal pulling, shoulder formation, shoulder rotation, equal diameter formation, and finishing.

[0003] The process begins with placing a quartz crucible onto a larger crucible, followed by the addition of polycrystalline silicon material. After loading, the furnace is closed, and a vacuum is created. Once the required vacuum is achieved, argon gas is introduced, and the heater is activated under slight negative pressure. The temperature of the molten material is gradually increased according to the process requirements until the desired melting temperature is reached, melting the solid polycrystalline silicon material into a liquid state. To maintain the stability of the melt and temperature field, and to ensure continuous crystal growth, polycrystalline silicon material needs to be continuously added to the quartz crucible within the furnace. After the polycrystalline silicon material has melted, crystal pulling cannot begin immediately because the temperature is higher than the crystal pulling temperature. Temperature adjustment is necessary to bring the temperature down to the optimal level for crystal pulling. Crystal pulling involves placing a seed crystal (a single crystal shaped to a specific form) pre-attached to the end of a steel wire rope into contact with the liquid surface. At the crystal pulling temperature, silicon molecules grow along the lattice direction of the seed crystal, thus forming a single crystal. Shoulder formation involves gradually growing the crystal to the required diameter. During shoulder formation, a section of crystal is pulled out, gradually increasing in length and diameter to approximately the desired diameter, to eliminate dislocations. Once the crystal has grown to the required diameter during shoulder formation, it enters the shoulder turning process. Shoulder turning controls the crystal diameter to the required production diameter. After shoulder turning is complete, the constant diameter control step begins. In this step, through automatic control of oxygen content and temperature, the crystal grows to a constant diameter as set.

[0004] The current Czochralski process, particularly during the melting and feeding stages, primarily relies on high-temperature duration and the contact area between the crucible and the molten polycrystalline silicon material, leading to the generation of a large amount of oxygen during crucible melting. Throughout the automated crystal pulling process, from the temperature adjustment stage to the shoulder-forming stage, considering oxygen generation and diffusion, it is also an enrichment process. If the amount of oxygen volatilized is greater than the amount generated, the oxygen content in the constant-diameter head stage (e.g., from the initial constant-diameter stage to when the crystal length increases by 50 mm) will be lower; conversely, the oxygen content will increase, resulting in a higher oxygen content in the constant-diameter head stage. The oxygen content in the constant-diameter head stage is difficult to predict accurately, making it difficult to pre-adjust the oxygen content and leading to unstable crystal quality. Summary of the Invention

[0005] In view of the above problems, embodiments of the present invention are proposed to provide an oxygen content prediction method that overcomes or at least partially solves the above problems, in order to address the difficulty in accurately predicting the oxygen content of crystals in the head stage of constant diameter, which in turn makes it difficult to adjust the oxygen content in advance and results in unstable crystal quality.

[0006] Accordingly, embodiments of the present invention also provide an oxygen content prediction device, an electronic device, and a storage medium to ensure the implementation and application of the above method.

[0007] To address the above problems, embodiments of the present invention disclose an oxygen content prediction method, comprising:

[0008] In this Czochralski single crystal growing process, characteristic data prior to the head stage of constant diameter growth were obtained, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth.

[0009] The feature data is input into the oxygen content prediction model, wherein the oxygen content prediction model is trained using feature data samples prior to the head stage of isodiameter growth and the oxygen content of samples from the corresponding labeled head stage of isodiameter growth.

[0010] Based on the characteristic data, the oxygen content prediction model generates the predicted oxygen content for the head stage of the isodiameter growth.

[0011] Optionally, the oxygen content prediction model includes multiple sub-oxygen content prediction models, and the step of inputting the feature data into the oxygen content prediction model includes:

[0012] Monitor the current treatment status before the head stage of isodiameter growth;

[0013] When the current processing state reaches a first preset processing state, the sub-oxygen content prediction model corresponding to the current processing state is invoked.

[0014] The feature data acquired before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state.

[0015] The step of generating the predicted oxygen content of the head stage of the isodiameter growth by the oxygen content prediction model based on the feature data includes:

[0016] Based on the characteristic data, the predicted oxygen content of the head stage of the isodiameter growth is generated by the sub-oxygen content prediction model under the current treatment state.

[0017] Optionally, the preset processing state includes at least one of the following: at least one processing state in the temperature adjustment stage where the time elapsed exceeds a first preset time elapsed; at least one processing state in the crystal pulling stage where the time elapsed exceeds a second preset time elapsed; at least one processing state in the shoulder forming stage where the time elapsed exceeds a third preset time elapsed; and at least one processing state in the shoulder turning stage where the time elapsed exceeds a fourth preset time elapsed.

[0018] Optionally, acquiring feature data prior to the head stage of isodiameter growth includes:

[0019] Monitor the current treatment status before the head stage of isodiameter growth;

[0020] When the current processing state reaches a second preset processing state, the running data is collected once, and the running data collected each time is added to the feature data.

[0021] Optionally, the oxygen content prediction model includes an integrated oxygen content prediction model. Before acquiring the characteristic data prior to the head stage of constant diameter growth during this Czochralski single crystal growth process, the method further includes:

[0022] The integrated oxygen content prediction model is obtained by integrating and learning multiple basic oxygen content prediction models.

[0023] Optionally, the feature data includes fused feature data, and the step of obtaining feature data prior to the head stage of isodiameter growth further includes:

[0024] Based on the correlation between different characteristic data during the Czochralski single crystal growth process, the fused characteristic data is generated according to multiple directly obtained operational data.

[0025] The directly acquired operational data and the fused feature data are used as the feature data input to the oxygen content prediction model.

[0026] Optionally, before inputting the feature data into the oxygen content prediction model, the method further includes:

[0027] After training the oxygen content prediction model using feature data samples prior to the head stage of isodiameter growth and the oxygen content samples of the corresponding labeled head stage of isodiameter growth, the feature weights corresponding to various feature data in the trained oxygen content prediction model are obtained.

[0028] After generating the predicted oxygen content for the head stage of the isodiameter growth based on the feature data using the oxygen content prediction model, the method further includes:

[0029] Based on the predicted oxygen content and the feature weights, the process parameters corresponding to each feature data are adjusted.

[0030] This invention also discloses an oxygen content prediction device, comprising:

[0031] The data acquisition module is used to acquire characteristic data before the head stage of constant diameter growth during this Czochralski single crystal growth process, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth.

[0032] The data input module is used to input the feature data into the oxygen content prediction model, wherein the oxygen content prediction model is trained by feature data samples before the head stage of isodiameter growth and the oxygen content of samples in the corresponding labeled head stage of isodiameter growth.

[0033] An oxygen content generation module is used to generate a predicted oxygen content for the head stage of the isodiameter growth based on the feature data and the oxygen content prediction model.

[0034] Optionally, the oxygen content prediction model includes multiple sub-oxygen content prediction models, and the data input module includes:

[0035] The status monitoring submodule is used to monitor the current processing status before the head stage of constant diameter growth;

[0036] The model invocation submodule is used to invoke the sub-oxygen content prediction model corresponding to the current processing state whenever the current processing state reaches a first preset processing state.

[0037] The feature data acquired before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state.

[0038] The oxygen content generation module includes:

[0039] An oxygen content generation submodule is used to generate, based on the feature data, a predicted oxygen content for the head stage of the isodiameter growth under the current processing state using the sub-oxygen content prediction model.

[0040] Optionally, the preset processing state includes at least one of the following: at least one processing state in the temperature adjustment stage where the time elapsed exceeds a first preset time elapsed; at least one processing state in the crystal pulling stage where the time elapsed exceeds a second preset time elapsed; at least one processing state in the shoulder forming stage where the time elapsed exceeds a third preset time elapsed; and at least one processing state in the shoulder turning stage where the time elapsed exceeds a fourth preset time elapsed.

[0041] Optionally, the data acquisition module includes:

[0042] The status monitoring submodule is used to monitor the current processing status before the head stage of constant diameter growth;

[0043] The data acquisition submodule is used to acquire the running data once each time the current processing state reaches a second preset processing state, and add the acquired running data to the feature data.

[0044] Optionally, the oxygen content prediction model includes an integrated oxygen content prediction model, and the device further includes:

[0045] An integrated learning module is used to integrate and learn multiple basic oxygen content prediction models before acquiring feature data before the head stage of constant diameter growth during this Czochralski single crystal growth process, to obtain the integrated oxygen content prediction model.

[0046] Optionally, the feature data includes fused feature data, and the data acquisition module further includes:

[0047] The data generation submodule is used to generate the fused feature data based on the correlation between different feature data during the Czochralski single crystal growth process and the multiple directly acquired operational data.

[0048] The data acquisition submodule is used to take the directly acquired running data and the fused feature data as the feature data input to the oxygen content prediction model.

[0049] Optionally, the device further includes:

[0050] The weight acquisition module is used to acquire the feature weights corresponding to various feature data in the trained oxygen content prediction model before the feature data is input into the oxygen content prediction model, after training the oxygen content prediction model by using feature data samples before the head stage of isodiameter growth and sample oxygen content of the corresponding labeled head stage of isodiameter growth.

[0051] The device further includes:

[0052] The parameter adjustment module is used to adjust the process parameters corresponding to each feature data according to the predicted oxygen content and the feature weights after the oxygen content prediction model generates the predicted oxygen content for the head stage of the isodiameter growth based on the feature data.

[0053] This invention also discloses an electronic device, characterized in that it includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0054] Memory, used to store computer programs;

[0055] When a processor executes a program stored in memory, it implements the method steps described above.

[0056] This invention also discloses a readable storage medium, which, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to execute one or more of the oxygen content prediction methods described in this invention.

[0057] The embodiments of the present invention have the following advantages:

[0058] According to an embodiment of the present invention, during the Czochralski single crystal growth process, characteristic data prior to the head stage of constant-diameter growth is acquired. This characteristic data includes operational data related to the oxygen content of the head stage of constant-diameter growth. The characteristic data is input into an oxygen content prediction model, which is trained using characteristic data samples prior to the head stage of constant-diameter growth and corresponding labeled oxygen content samples from the head stage. Based on the characteristic data, the oxygen content prediction model generates a predicted oxygen content for the head stage of constant-diameter growth. This allows the oxygen content of the head stage of constant-diameter growth to be predicted in advance, thereby enabling pre-control of the oxygen content and improving crystal quality. Attached Figure Description

[0059] Figure 1 This is a flowchart illustrating the steps of an embodiment of the oxygen content prediction method of the present invention;

[0060] Figure 2 This is a flowchart illustrating the steps of an embodiment of the oxygen content prediction method of the present invention;

[0061] Figure 3 This is a schematic diagram illustrating an example of oxygen content prediction;

[0062] Figure 4 This is a structural block diagram of an embodiment of an oxygen content prediction device of the present invention;

[0063] Figure 5 This is a structural block diagram of a computing device for predicting oxygen content, according to an exemplary embodiment. Detailed Implementation

[0064] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0065] Reference Figure 1 The diagram illustrates a flowchart of an embodiment of the oxygen content prediction method of the present invention, which may specifically include the following steps:

[0066] Step 101: In this Czochralski single crystal growth process, acquire characteristic data prior to the head stage of constant diameter growth, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth.

[0067] In this embodiment of the invention, the Czochralski single crystal process is the process of refining raw materials into single crystals using the Czochralski method, such as the process of Czochralski single crystal silicon. The Czochralski single crystal process can be divided into the melting stage, the feeding stage, the temperature control stage, the crystal pulling stage, the shoulder formation stage, the shoulder turning stage, and the constant diameter growth stage. Among them, the part at the beginning of the constant diameter growth stage is called the head stage of constant diameter growth (for example, the stage from the start of constant diameter growth to the crystal length increasing by 50 mm).

[0068] In this embodiment of the invention, the Czochralski single crystal growing process includes operational data such as power, crucible rotation, and process time. The operational data includes data monitored during actual operation and data set for the equipment; specifically, it can include any applicable data, and this embodiment of the invention does not impose any limitations on this. Through extensive data analysis, it was found that the oxygen content in the head stage of constant-diameter growth is related to some operational data prior to the head stage. Through data analysis, operational data related to the oxygen content in the head stage of constant-diameter growth were selected from the operational data prior to the head stage and used as feature data.

[0069] For example, based on the mechanism of oxygen content in the head stage of isodiameter growth and historical data analysis, operational data related to oxygen content in the head stage of isodiameter growth are first screened. Then, data analysis and processing are performed. A second screening is conducted, performing correlation analysis between operational data points and between operational data and oxygen content, ultimately determining the characteristic data that can be used for data modeling.

[0070] Step 102: Input the feature data into the oxygen content prediction model, wherein the oxygen content prediction model is trained using feature data samples prior to the head stage of isodiameter growth and the oxygen content samples of the corresponding labeled head stage of isodiameter growth.

[0071] In this embodiment of the invention, the oxygen content of the head stage of isodiameter growth can be predicted by machine learning of the correlation between feature data prior to the head stage and the oxygen content of the head stage, thus obtaining an oxygen content prediction model. To train the oxygen content prediction model, sample data and corresponding label data are needed, namely, feature data samples prior to the head stage of isodiameter growth and the corresponding labeled oxygen content samples of the head stage. Specifically, feature data samples and sample oxygen contents can be obtained through multiple experiments, or feature data samples and sample oxygen contents can be selected from historical data.

[0072] In an optional embodiment of the present invention, the oxygen content prediction model includes: a Support Vector Machine (SVM) model, a K-NN (K-Nearest Neighbor) model, a Naive Bayes model, a Random Forest model, an extreme gradient boosting model, or a class gradient boosting model. The SVM model constructs a hyperplane or set of hyperplanes in a high-dimensional or infinite-dimensional space, which can be used for classification, regression, or other tasks. The K-NN model classifies by measuring the distance between different feature values, meaning that each sample can be represented by its K nearest neighbors. The Naive Bayes model is a simple method for building a classifier that assigns a class label, represented by feature values, to problem instances, with the class labels taken from a finite set. The Random Forest model is a classifier that uses multiple trees to train and predict samples. The extreme gradient boosting (XGboost) model is a model that uses boosting trees for prediction in massive parallel. Categorical Gradient Boosting (CATboost) is a gradient boosting algorithm that can handle categorical features very well.

[0073] In this embodiment of the invention, during the prediction stage, the input to the oxygen content prediction model is the feature data prior to the head stage of isodiameter growth, and the output is the predicted value of oxygen content, denoted as the predicted oxygen content.

[0074] Step 103: Based on the feature data, generate the predicted oxygen content of the head stage of the isodiameter growth using the oxygen content prediction model.

[0075] In this embodiment of the invention, for the current Czochralski crystal growing process, after obtaining the current feature data, the obtained feature data is input into the oxygen content prediction model to obtain the output of the oxygen content prediction model, and then the output of the oxygen content prediction model is used as the predicted oxygen content of the head stage of constant diameter growth.

[0076] In this embodiment of the invention, before the head stage of constant diameter growth, the characteristic data at that time can be input into the oxygen content prediction model after a preset time period or after the process reaches a first preset processing state, so as to continuously obtain updated predicted oxygen content.

[0077] According to an embodiment of the present invention, during the Czochralski single crystal growth process, characteristic data prior to the head stage of constant-diameter growth is acquired. This characteristic data includes operational data related to the oxygen content of the head stage of constant-diameter growth. The characteristic data is input into an oxygen content prediction model, which is trained using characteristic data samples prior to the head stage of constant-diameter growth and corresponding labeled oxygen content samples from the head stage. Based on the characteristic data, the oxygen content prediction model generates a predicted oxygen content for the head stage of constant-diameter growth. This allows the oxygen content of the head stage of constant-diameter growth to be predicted in advance, thereby enabling pre-control of the oxygen content and improving crystal quality.

[0078] In an optional embodiment of the present invention, in a specific implementation, before inputting the feature data into the oxygen content prediction model, the method may further include: after training the oxygen content prediction model using feature data samples prior to the head stage of constant diameter growth and the oxygen content of samples from the corresponding labeled head stage of constant diameter growth, obtaining the feature weights corresponding to various feature data in the trained oxygen content prediction model; and after generating the predicted oxygen content for the head stage of constant diameter growth from the oxygen content prediction model based on the feature data, the method may further include: adjusting the process parameters corresponding to each feature data based on the predicted oxygen content and the feature weights.

[0079] The training process of an oxygen content prediction model involves determining the feature weights corresponding to various feature data based on the training data. After the oxygen content prediction model is trained, the feature weights are determined. The feature weights corresponding to various feature data are then extracted from the trained oxygen content prediction model. The larger the feature weight, the greater the impact of that feature data on oxygen content; conversely, the smaller the feature weight, the smaller the impact of that feature data on oxygen content.

[0080] Then, after generating the predicted oxygen content, the process parameters corresponding to each feature data are adjusted based on the predicted oxygen content and feature weights. Since the change in oxygen content is delayed when the relevant process parameters are changed to control it, the oxygen content needs to be controlled before the head stage of constant diameter growth. The process parameters, corresponding to the feature data, are process-related parameters set for the equipment, including power, crucible rotation, etc., or any other applicable process parameters; this embodiment of the invention does not impose any limitations on this.

[0081] The operating data includes at least one of the following: power and crucible rotation; the process parameters include at least one of the following: power and crucible rotation. Power refers to the power of the single crystal furnace. Crucible rotation is the rotational speed at which the melt in the crucible of the single crystal furnace is stirred.

[0082] In this embodiment of the invention, if the predicted oxygen content is too high, the process parameters are adjusted to lower the oxygen content, bringing it closer to a suitable level. Each process parameter and its corresponding feature data has a corresponding relationship. Based on the predicted oxygen content and the feature weights corresponding to each process parameter, the adjustment amount for each process parameter is calculated. For example, for each process parameter, the difference between the predicted oxygen content and the target oxygen content is multiplied by the feature weight corresponding to the process parameter, and the product is determined as the adjustment amount for that process parameter. Any applicable method for calculating the adjustment amount can be used, and this embodiment of the invention does not impose any limitations on this. Finally, based on the adjustment amount for each process parameter, the parameters are increased or decreased.

[0083] In one optional embodiment of the present invention, in a specific implementation, the oxygen content prediction model includes an integrated oxygen content prediction model. Before obtaining the feature data before the head stage of constant diameter growth during the current Czochralski single crystal growth process, the method may further include: integrating and learning multiple basic oxygen content prediction models to obtain the integrated oxygen content prediction model.

[0084] Multiple suitable models are selected as the base models for ensemble learning, generally an odd number (e.g., 3 or 5). For the training set data, by training multiple basic oxygen content prediction models and employing a certain combination strategy, an ensemble oxygen content prediction model can be ultimately formed, achieving the goal of leveraging the strengths of multiple models. The ensemble learning algorithm can employ Bagging (Bootstrap aggregating), Boosting, Stacking, MOPSO (Multiple Objective Particle Swarm Optimization), or any other applicable method; this embodiment of the invention does not impose any limitations on these methods.

[0085] In an optional embodiment of the present invention, the feature data includes fused feature data. In a specific implementation of obtaining the feature data before the head stage of equal diameter growth, it may further include: based on the correlation between different feature data in the Czochralski single crystal growth process, generating the fused feature data according to the multiple directly obtained operating data, and using the directly obtained operating data and the fused feature data as feature data input to the oxygen content prediction model.

[0086] In the Czochralski (CZ) single crystal growth process, some characteristic data can be directly obtained from operational data. However, some characteristic data that cannot be directly obtained can be obtained through feature fusion. Based on the correlation between different characteristic data in the CZ single crystal growth process, multiple directly obtained operational data are fused to generate fused characteristic data. For example, the weighted sum of directly obtained power and crucible rotation data yields a fused characteristic data. Because the fused characteristic data not only contains multiple operational data but also information on the correlation between these data, the oxygen content prediction model can more accurately predict oxygen content, improving the model's accuracy.

[0087] Reference Figure 2 The diagram illustrates a flowchart of an embodiment of the oxygen content prediction method of the present invention, which may specifically include the following steps:

[0088] Step 201: In this Czochralski single crystal growth process, acquire characteristic data prior to the head stage of constant diameter growth, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth.

[0089] Step 202: Monitor the current processing status before the head stage of isodiameter growth.

[0090] In this embodiment of the invention, before the head stage of constant diameter growth, the processing state of the process includes the state of crystal temperature, crystal length, crystal diameter, etc., or any other applicable processing state. This embodiment of the invention does not limit this.

[0091] In this embodiment of the invention, the processing status of the process is monitored before the head stage of constant diameter growth to obtain the current processing status. For example, the crystal length is monitored to obtain the current crystal length.

[0092] Step 203: When the current processing state reaches a first preset processing state, the sub-oxygen content prediction model corresponding to the current processing state is invoked.

[0093] In this embodiment of the invention, the relationship between feature data and oxygen content differs at different stages. Therefore, different models need to be trained for prediction. The oxygen content prediction model includes multiple sub-oxygen content prediction models. After determining a first preset processing state, the corresponding sub-oxygen content prediction model can be obtained. When the current processing state reaches a certain first preset processing state, the sub-oxygen content prediction model corresponding to that first preset processing state is invoked.

[0094] For example, the trained oxygen content prediction model is deployed to an industrial computer and a PAC (Programmable Automation Controller) system. Four preset prediction points are called sequentially to predict the results, achieving the goal of predicting the oxygen content in the head stage of constant-diameter growth. Preferably, after model training, edge computing is used for deployment to the PAC for prediction. Since the PAC is closest to the equipment in this Czochralski single crystal growth process, the delay in oxygen content prediction is reduced.

[0095] In an optional embodiment of the present invention, the preset processing state includes at least one of the following: at least one processing state in the temperature adjustment stage where the time between the end and the end of the process exceeds a first preset time; at least one processing state in the crystal pulling stage where the time between the end and the end of the process exceeds a second preset time; at least one processing state in the shoulder forming stage where the time between the end and the end of the process exceeds a third preset time; and at least one processing state in the shoulder turning stage where the time between the end and the end of the process exceeds a fourth preset time.

[0096] Because the change in oxygen content is delayed when controlling related process parameters, it is necessary to control the process at a certain point in time before the end of each stage (exceeding a preset time interval) to ensure that the adjustment of process parameters is reflected in the oxygen content later. The first, second, third, and fourth preset time intervals can be set according to actual needs, and this embodiment of the invention does not impose any restrictions on them. At least one processing state in the temperature adjustment stage that exceeds the first preset time interval includes the processing state where the temperature reaches a preset temperature, or any other applicable processing state, and this embodiment of the invention does not impose any restrictions on it. At least one processing state in the crystal pulling stage that exceeds the second preset time interval includes the processing state where the crystal length reaches a preset length, or any other applicable processing state, and this embodiment of the invention does not impose any restrictions on it. At least one processing state in the shoulder forming stage that exceeds the third preset time interval includes the processing state where the crystal length reaches a preset length, or any other applicable processing state, and this embodiment of the invention does not impose any restrictions on it. At least one processing state in the shoulder turning stage that is more than four preset time intervals from the end of the shoulder turning stage includes the processing state in which the crystal diameter reaches a preset diameter, or any other applicable processing state. This embodiment of the invention does not limit this.

[0097] Step 204: Input the feature data obtained before the current processing state into the sub-oxygen content prediction model corresponding to the current processing state.

[0098] In this embodiment of the invention, the feature data is collected multiple times during the Czochralski single crystal growth process. The feature data acquired once or multiple times before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state, so as to obtain the predicted oxygen content under the current processing state through the sub-oxygen content prediction model.

[0099] In an optional embodiment of the present invention, a specific implementation of acquiring feature data prior to the head stage of equal diameter growth includes: monitoring the current processing state prior to the head stage of equal diameter growth; and, each time the current processing state reaches a second preset processing state, collecting the running data once, and adding the collected running data to the feature data.

[0100] To predict oxygen content multiple times based on characteristic data from a period preceding the head stage of constant-diameter growth, operational data needs to be collected at regular intervals. Therefore, multiple data collection points are set up during the process preceding the head stage of constant-diameter growth, each corresponding to a second preset processing state in the process. For example, this could be the temperature reaching a preset value during the temperature adjustment stage, the crystal length reaching a preset value during the crystal pulling stage, the crystal length reaching a preset value during the shoulder formation stage, or the crystal diameter reaching a preset value during the shoulder turning stage. Specifically, any applicable second preset processing state can be included, and this embodiment of the invention does not limit this.

[0101] Each time the current processing state reaches a second preset processing state, the operating data is collected once. The collected operating data is then added to the feature data. For example, the melting stage, feeding stage, temperature adjustment stage, crystal pulling stage, shoulder forming stage, and shoulder turning stage are divided into 17 collection points, i.e., 17 second preset processing states. Each collection point collects 8 basic feature data, such as power and crucible rotation, as well as 1 time-related feature data, such as the time taken for a certain process.

[0102] Step 205: Based on the feature data, generate the predicted oxygen content of the head stage of the isodiameter growth under the current treatment state using the sub-oxygen content prediction model.

[0103] In this embodiment of the invention, based on the feature data obtained before the current processing state, the sub-oxygen content prediction model can generate a predicted oxygen content, that is, the predicted oxygen content of the head stage of isodiameter growth under the current processing state.

[0104] According to an embodiment of the present invention, during the Czochralski single crystal growth process, characteristic data prior to the head stage of constant diameter growth is acquired. This characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth. The current processing state prior to the head stage of constant diameter growth is monitored. Whenever the current processing state reaches a first preset processing state, the sub-oxygen content prediction model corresponding to the current processing state is invoked. The characteristic data acquired prior to the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state. Based on the characteristic data, the sub-oxygen content prediction model generates a predicted oxygen content for the head stage of constant diameter growth under the current processing state. This allows the oxygen content of the head stage of constant diameter growth to be predicted in advance, thereby enabling pre-control of the oxygen content and improving crystal quality.

[0105] To enable those skilled in the art to better understand this application, the following specific examples illustrate one implementation of this application.

[0106] like Figure 3 The diagram illustrates one example of oxygen content prediction, which may specifically include the following steps.

[0107] Data acquisition: Data acquisition for the Czochralski single crystal growth process is performed through the system and central control platform.

[0108] Data preprocessing: Data preprocessing is performed on the collected data, mainly to detect outliers and filter out useless data.

[0109] Feature engineering involves extracting input feature data, selecting features beneficial for oxygen content prediction, and generating new features through feature fusion. This reduces feature dimensionality and combines existing features to generate fused features, which are then used as feature data samples.

[0110] Model selection: Select appropriate models as the base models for ensemble learning, generally an odd number of models (3 or 5).

[0111] Model fusion: Selected candidate models are fused using ensemble learning algorithms such as Bagging, Boosting, Stacking, and MOPSO to obtain the final model.

[0112] Model training: Generally, hold-out or cross-validation methods are used for sample classification. With hold-out, the training set typically occupies 2 / 3 or 4 / 5 of the data, with the remainder being the test set. Model training is then performed after the data is partitioned.

[0113] Model deployment: After the model is trained, it is deployed using edge computing to the system PAC for prediction.

[0114] Multiple predictions: The model was called up at the temperature adjustment stage, crystal introduction stage, shoulder formation stage, and shoulder transformation stage to predict oxygen content.

[0115] Core features: Sort the feature data by weight and output the feature data with the core influence.

[0116] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0117] Reference Figure 4 The diagram shows a structural block diagram of an embodiment of an oxygen content prediction device according to the present invention, which may specifically include the following modules:

[0118] The data acquisition module 301 is used to acquire characteristic data before the head stage of constant diameter growth during this Czochralski single crystal growth process, wherein the characteristic data includes operating data related to the oxygen content of the head stage of constant diameter growth.

[0119] The data input module 302 is used to input the feature data into the oxygen content prediction model, wherein the oxygen content prediction model is trained by feature data samples before the head stage of isodiameter growth and the oxygen content of samples in the corresponding labeled head stage of isodiameter growth.

[0120] The oxygen content generation module 303 is used to generate the predicted oxygen content of the head stage of the isodiameter growth by the oxygen content prediction model based on the feature data.

[0121] Optionally, the oxygen content prediction model includes multiple sub-oxygen content prediction models, and the data input module includes:

[0122] The status monitoring submodule is used to monitor the current processing status before the head stage of constant diameter growth;

[0123] The model invocation submodule is used to invoke the sub-oxygen content prediction model corresponding to the current processing state whenever the current processing state reaches a first preset processing state.

[0124] The feature data acquired before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state.

[0125] The oxygen content generation module includes:

[0126] An oxygen content generation submodule is used to generate, based on the feature data, a predicted oxygen content for the head stage of the isodiameter growth under the current processing state using the sub-oxygen content prediction model.

[0127] Optionally, the preset processing state includes at least one of the following: at least one processing state in the temperature adjustment stage where the time elapsed exceeds a first preset time elapsed; at least one processing state in the crystal pulling stage where the time elapsed exceeds a second preset time elapsed; at least one processing state in the shoulder forming stage where the time elapsed exceeds a third preset time elapsed; and at least one processing state in the shoulder turning stage where the time elapsed exceeds a fourth preset time elapsed.

[0128] Optionally, the data acquisition module includes:

[0129] The status monitoring submodule is used to monitor the current processing status before the head stage of constant diameter growth;

[0130] The data acquisition submodule is used to acquire the running data once each time the current processing state reaches a second preset processing state, and add the acquired running data to the feature data.

[0131] Optionally, the oxygen content prediction model includes an integrated oxygen content prediction model, and the device further includes:

[0132] An integrated learning module is used to integrate and learn multiple basic oxygen content prediction models before acquiring feature data before the head stage of constant diameter growth during this Czochralski single crystal growth process, to obtain the integrated oxygen content prediction model.

[0133] Optionally, the feature data includes fused feature data, and the data acquisition module further includes:

[0134] The data generation submodule is used to generate the fused feature data based on the correlation between different feature data during the Czochralski single crystal growth process and the multiple directly acquired operational data.

[0135] The data acquisition submodule is used to take the directly acquired running data and the fused feature data as the feature data input to the oxygen content prediction model.

[0136] Optionally, the device further includes:

[0137] The weight acquisition module is used to acquire the feature weights corresponding to various feature data in the trained oxygen content prediction model before the feature data is input into the oxygen content prediction model, after training the oxygen content prediction model by using feature data samples before the head stage of isodiameter growth and sample oxygen content of the corresponding labeled head stage of isodiameter growth.

[0138] The device further includes:

[0139] The parameter adjustment module is used to adjust the process parameters corresponding to each feature data according to the predicted oxygen content and the feature weights after the oxygen content prediction model generates the predicted oxygen content for the head stage of the isodiameter growth based on the feature data.

[0140] According to an embodiment of the present invention, during the Czochralski single crystal growth process, characteristic data prior to the head stage of constant-diameter growth is acquired. This characteristic data includes operational data related to the oxygen content of the head stage of constant-diameter growth. The characteristic data is input into an oxygen content prediction model, which is trained using characteristic data samples prior to the head stage of constant-diameter growth and corresponding labeled oxygen content samples from the head stage. Based on the characteristic data, the oxygen content prediction model generates a predicted oxygen content for the head stage of constant-diameter growth. This allows the oxygen content of the head stage of constant-diameter growth to be predicted in advance, thereby enabling pre-control of the oxygen content and improving crystal quality.

[0141] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0142] Figure 5 This is a structural block diagram illustrating an electronic device 400 for oxygen content prediction according to an exemplary embodiment. For example, the electronic device 400 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.

[0143] Reference Figure 5 The electronic device 400 may include one or more of the following components: processing component 402, memory 404, power supply component 406, multimedia component 408, audio component 410, input / output (I / O) interface 412, sensor component 414, and communication component 416.

[0144] Processing component 402 typically controls the overall operation of electronic device 400, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 402 may include one or more processors 420 to execute instructions to complete all or part of the steps of the oxygen content prediction method described above. Furthermore, processing component 402 may include one or more modules to facilitate interaction between processing component 402 and other components. For example, processing component 402 may include a multimedia module to facilitate interaction between multimedia component 408 and processing component 402.

[0145] Memory 404 is configured to store various types of data to support the operation of device 400. Examples of this data include instructions for any application or method operating on electronic device 400, contact data, phonebook data, messages, pictures, videos, etc. Memory 404 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0146] Power component 404 provides power to various components of electronic device 400. Power component 404 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 400.

[0147] Multimedia component 408 includes a screen that provides an output interface between the electronic device 400 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 408 includes a front-facing camera and / or a rear-facing camera. When the electronic device 400 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0148] Audio component 410 is configured to output and / or input audio signals. For example, audio component 410 includes a microphone (MIC) configured to receive external audio signals when electronic device 400 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 404 or transmitted via communication component 416. In some embodiments, audio component 410 also includes a speaker for outputting audio signals.

[0149] I / O interface 412 provides an interface between processing component 402 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0150] Sensor assembly 414 includes one or more sensors for providing state assessments of various aspects of electronic device 400. For example, sensor assembly 414 may detect the on / off state of device 400, the relative positioning of components such as the display and keypad of electronic device 400, changes in position of electronic device 400 or a component of electronic device 400, the presence or absence of user contact with electronic device 400, orientation or acceleration / deceleration of electronic device 400, and temperature changes of electronic device 400. Sensor assembly 414 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 414 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 414 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0151] Communication component 416 is configured to facilitate wired or wireless communication between electronic device 400 and other devices. Electronic device 400 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 414 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 414 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0152] In an exemplary embodiment, the electronic device 400 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the oxygen content prediction method described above.

[0153] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 404 including instructions, which can be executed by a processor 420 of an electronic device 400 to complete the oxygen content prediction method described above. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0154] A non-transitory computer-readable storage medium, when instructions in the storage medium are executed by a terminal's processor, enables the terminal to perform an oxygen content prediction method, the method comprising:

[0155] In this Czochralski single crystal growing process, characteristic data prior to the head stage of constant diameter growth were obtained, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth.

[0156] The feature data is input into the oxygen content prediction model, wherein the oxygen content prediction model is trained using feature data samples prior to the head stage of isodiameter growth and the oxygen content of samples from the corresponding labeled head stage of isodiameter growth.

[0157] Based on the characteristic data, the oxygen content prediction model generates the predicted oxygen content for the head stage of the isodiameter growth.

[0158] Optionally, the oxygen content prediction model includes multiple sub-oxygen content prediction models, and the step of inputting the feature data into the oxygen content prediction model includes:

[0159] Monitor the current treatment status before the head stage of isodiameter growth;

[0160] When the current processing state reaches a first preset processing state, the sub-oxygen content prediction model corresponding to the current processing state is invoked.

[0161] The feature data acquired before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state.

[0162] The step of generating the predicted oxygen content of the head stage of the isodiameter growth by the oxygen content prediction model based on the feature data includes:

[0163] Based on the characteristic data, the predicted oxygen content of the head stage of the isodiameter growth is generated by the sub-oxygen content prediction model under the current treatment state.

[0164] Optionally, the preset processing state includes at least one of the following: at least one processing state in the temperature adjustment stage where the time elapsed exceeds a first preset time elapsed; at least one processing state in the crystal pulling stage where the time elapsed exceeds a second preset time elapsed; at least one processing state in the shoulder forming stage where the time elapsed exceeds a third preset time elapsed; and at least one processing state in the shoulder turning stage where the time elapsed exceeds a fourth preset time elapsed.

[0165] Optionally, acquiring feature data prior to the head stage of isodiameter growth includes:

[0166] Monitor the current treatment status before the head stage of isodiameter growth;

[0167] When the current processing state reaches a second preset processing state, the running data is collected once, and the running data collected each time is added to the feature data.

[0168] Optionally, the oxygen content prediction model includes an integrated oxygen content prediction model. Before acquiring the characteristic data prior to the head stage of constant diameter growth during this Czochralski single crystal growth process, the method further includes:

[0169] The integrated oxygen content prediction model is obtained by integrating and learning multiple basic oxygen content prediction models.

[0170] Optionally, the feature data includes fused feature data, and the step of obtaining feature data prior to the head stage of isodiameter growth further includes:

[0171] Based on the correlation between different characteristic data during the Czochralski single crystal growth process, the fused characteristic data is generated according to multiple directly obtained operational data.

[0172] The directly acquired operational data and the fused feature data are used as the feature data input to the oxygen content prediction model.

[0173] Optionally, before inputting the feature data into the oxygen content prediction model, the method further includes:

[0174] After training the oxygen content prediction model using feature data samples prior to the head stage of isodiameter growth and the oxygen content samples of the corresponding labeled head stage of isodiameter growth, the feature weights corresponding to various feature data in the trained oxygen content prediction model are obtained.

[0175] After generating the predicted oxygen content for the head stage of the isodiameter growth based on the feature data using the oxygen content prediction model, the method further includes:

[0176] Based on the predicted oxygen content and the feature weights, the process parameters corresponding to each feature data are adjusted.

[0177] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0178] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0179] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0180] These computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing terminal device to operate in a predictive manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0181] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0182] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0183] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0184] The above provides a detailed description of an oxygen content prediction method and apparatus, an electronic device, and a storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for predicting oxygen content, characterized in that, include: In this Czochralski single crystal growing process, characteristic data prior to the head stage of constant diameter growth were obtained, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth. The feature data is input into the oxygen content prediction model, wherein the oxygen content prediction model is trained using feature data samples prior to the head stage of isodiameter growth and the oxygen content of samples from the corresponding labeled head stage of isodiameter growth. Based on the characteristic data, the oxygen content prediction model generates the predicted oxygen content of the head stage of the isodiameter growth. The oxygen content prediction model includes an integrated oxygen content prediction model. Before acquiring the characteristic data prior to the head stage of constant diameter growth during this Czochralski single crystal growth process, the method further includes: The integrated oxygen content prediction model is obtained by integrating and learning multiple basic oxygen content prediction models.

2. The method according to claim 1, characterized in that, The oxygen content prediction model includes multiple sub-oxygen content prediction models, and the step of inputting the feature data into the oxygen content prediction model includes: Monitor the current treatment status before the head stage of isodiameter growth; When the current processing state reaches a first preset processing state, the sub-oxygen content prediction model corresponding to the current processing state is invoked. The feature data acquired before the current processing state is input into the sub-oxygen content prediction model corresponding to the current processing state. The step of generating the predicted oxygen content of the head stage of the isodiameter growth by the oxygen content prediction model based on the feature data includes: Based on the characteristic data, the predicted oxygen content of the head stage of the isodiameter growth is generated by the sub-oxygen content prediction model under the current treatment state.

3. The method according to claim 2, characterized in that, The preset processing state includes at least one of the following: at least one processing state in which the temperature adjustment stage is more than a first preset time after the end; at least one processing state in which the crystal pulling stage is more than a second preset time after the end; at least one processing state in which the shoulder forming stage is more than a third preset time after the end; and at least one processing state in which the shoulder turning stage is more than a fourth preset time after the end.

4. The method according to claim 2, characterized in that, The acquisition of feature data prior to the head stage of isodiameter growth includes: Monitor the current treatment status before the head stage of isodiameter growth; When the current processing state reaches a second preset processing state, the running data is collected once, and the running data collected each time is added to the feature data.

5. The method according to claim 1, characterized in that, The feature data includes the fused feature data, and the acquisition of feature data prior to the head stage of isodiameter growth further includes: Based on the correlation between different characteristic data during the Czochralski single crystal growth process, the fused characteristic data is generated according to multiple directly obtained operational data. The directly acquired operational data and the fused feature data are used as the feature data input to the oxygen content prediction model.

6. The method according to claim 1, characterized in that, Before inputting the feature data into the oxygen content prediction model, the method further includes: After training the oxygen content prediction model using feature data samples prior to the head stage of isodiameter growth and the oxygen content samples of the corresponding labeled head stage of isodiameter growth, the feature weights corresponding to various feature data in the trained oxygen content prediction model are obtained. After generating the predicted oxygen content for the head stage of the isodiameter growth based on the feature data using the oxygen content prediction model, the method further includes: Based on the predicted oxygen content and the feature weights, the process parameters corresponding to each feature data are adjusted.

7. An oxygen content prediction device, characterized in that, include: The data acquisition module is used to acquire characteristic data before the head stage of constant diameter growth during this Czochralski single crystal growth process, wherein the characteristic data includes operational data related to the oxygen content of the head stage of constant diameter growth. The data input module is used to input the feature data into the oxygen content prediction model, wherein the oxygen content prediction model is trained by feature data samples before the head stage of isodiameter growth and the oxygen content of samples in the corresponding labeled head stage of isodiameter growth. An oxygen content generation module is used to generate a predicted oxygen content for the head stage of the isodiameter growth based on the feature data and the oxygen content prediction model. The oxygen content prediction model includes an integrated oxygen content prediction model, and the device further includes: An integrated learning module is used to integrate and learn multiple basic oxygen content prediction models before acquiring feature data before the head stage of constant diameter growth during this Czochralski single crystal growth process, to obtain the integrated oxygen content prediction model.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-6.

9. A readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is able to perform one or more of the oxygen content prediction methods as described in claims 1-6.

Images