Automatic recommendation method and system for seed crystal pulling power and seed crystal pulling crucible position in crystal pulling industry, and medium

Abstract

Provided in the present invention are an automatic recommendation method for seed crystal pulling power and a seed crystal pulling crucible position in the crystal pulling industry, characterized in that the method comprises the following steps: step one, acquiring historical device parameters of a crystal pulling device in a historical seed crystal rotation stage and a historical seed crystal pulling stage; step two, inputting the historical device parameters into a machine learning model, performing training to optimize the model performance until the accuracy rate of the model reaches 95% or above, and then outputting the optimized model; step three, acquiring current-section device parameters of the crystal pulling device; and step four, inputting the current-section device parameters into the optimized model, and predicting the current seed crystal pulling power and seed crystal pulling crucible position.

Description

Method, system and medium for automatically recommending seeding power and seeding pot position in the crystal pulling industry Technical Field

[0001] The present invention relates to the field of semiconductor production technology, and in particular to a method, system and computer-readable storage medium for automatically recommending seeding power and seeding pot positions in the crystal pulling industry. Background Art

[0002] In the semiconductor crystal pulling manufacturing industry, it is currently necessary to rely on manual experience to calculate the appropriate parameter values ​​of the crystal seeding crucible position and crystal seeding power, and then input them into the equipment to seed the crystal. Due to differences in human experience, the crystal seeding success rate of the equipment fluctuates greatly, resulting in a low crystal seeding success rate.

[0003] Summary of the Invention

[0004] In order to solve the above problems, the present invention aims to provide a method, system and computer-readable storage medium for automatically recommending seeding power and seeding pot position in the crystal pulling industry.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] The present invention provides a method for automatically recommending seeding power and seeding pot position in the crystal pulling industry, which is characterized by comprising the following steps:

[0007] Step 1: Obtain historical equipment parameters of the crystal pulling equipment during the historical seed crystal rotation stage and seeding stage;

[0008] Step 2: Input historical device parameters into the machine learning model for training and optimization of model performance until the model accuracy reaches above 95%, and then output the optimized model;

[0009] Step 3: Obtain the current segment equipment parameters of the crystal pulling equipment;

[0010] Step 4: Input the current equipment parameters into the optimized model to predict the current seeding power and seeding pot position.

[0011] Furthermore, the method for automatically recommending seeding power and seeding pot position in the crystal pulling industry provided by the present invention may also have the following features: among the historical equipment parameters, the equipment parameters of the historical seed crystal rotation stage are all equipment parameters including batches, number of sections, and residual material weight, and the equipment parameters of the historical seeding stage are seeding power and seeding pot position.

[0012] Furthermore, the method for automatically recommending seeding power and seeding pot position in the crystal pulling industry provided by the present invention may also have the following feature: wherein the historical equipment parameters are equipment parameters within a historical year.

[0013] Furthermore, the automatic recommendation method for seeding power and seeding pot position in the crystal pulling industry provided by the present invention can also have the following features: wherein, the machine learning model adopts the XGboost regression model to divide the historical equipment parameters into a training set and a validation set. During the model training process, the independent variables are the equipment parameters of the historical seeding stage, and the dependent variables are the seeding power and seeding pot position of the historical seeding stage.

[0014] Furthermore, the method for automatically recommending seeding power and seeding pot position in the crystal pulling industry provided by the present invention may also have the following features: wherein, the current section equipment parameters of the crystal pulling equipment obtained in step three are all equipment parameters including batch, number of sections, and residual material weight, which are all actual values ​​of the current section.

[0015] The present invention also provides an automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry, which is characterized by including: a historical equipment parameter storage module, which stores the historical equipment parameters of the equipment in the historical seed crystal rotation stage and seeding stage according to the equipment name; a current section equipment parameter acquisition module, which is used to obtain the current section equipment parameters from the production line; a prediction module, which inputs the current section equipment parameters, calls the optimized model, and predicts the seeding power and seeding pot position in the seeding stage through model calculation; a display interface module, which is used to display the equipment parameter fields and parameter change trends in the system interface.

[0016] Furthermore, the automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry provided by the present invention may also have the following features: the equipment parameter fields displayed on the system interface include: equipment name, batch of the local segment, number of segments of the current segment, remaining weight of the current segment, predicted recommended seeding power, and predicted seeding pot position.

[0017] Furthermore, the automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry provided by the present invention may also have the following features: wherein the parameter change trend includes a table form and a trend graph form.

[0018] The present invention also provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the above-mentioned method for automatically recommending seeding power and seeding pot position in the crystal pulling industry is implemented.

[0019] Beneficial effects of the present invention:

[0020] The present invention provides a method and system for automatically recommending seeding power and seeding pot positions in the crystal pulling industry. This system uses machine learning to perform data mining on the historical seeding power and pot positions of the equipment, quickly and automatically recommends seeding power and pot position values, guides on-site production, improves the seeding success rate, and effectively reduces the burden on on-site personnel and the experience requirements of the personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG1 is a flow chart of a method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to an embodiment of the present invention;

[0022] FIG2 is a system interface in an embodiment of the present invention (the parameter change trend displayed below is in table form);

[0023] FIG3 is a system interface in an embodiment of the present invention (the parameter change trend displayed below is in the form of a trend graph);

[0024] FIG4 is an enlarged diagram of a trend diagram of seeding power and seeding pot position in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the following embodiments are combined with the accompanying drawings to specifically illustrate the technical solutions of the present invention.

[0026] An embodiment of the present invention provides a system for automatically recommending seeding power and seeding pot positions in the crystal pulling industry and a computer-readable storage medium. The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement a system for automatically recommending seeding power and seeding pot positions in the crystal pulling industry.

[0027] 1 , a method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to an embodiment of the present invention includes the following steps:

[0028] Step 1: Obtain historical equipment parameters for the crystal pulling equipment during the seeding and seeding stages. Historical equipment parameters for the seeding stage include all equipment parameters, including batch size, number of stages, and residual material weight. Historical equipment parameters for the seeding stage include seeding power and seeding pot position. Historical equipment parameters refer to equipment parameters from the past year.

[0029] Step 2: Input historical equipment parameters into the machine learning model for training and optimization, until the model accuracy reaches over 95%, and then output the optimized model. The machine learning model uses the XGboost regression model, and the historical equipment parameters are divided into a training set and a validation set. The model training process uses the training set for training and the validation set for accuracy verification. The independent variables are the equipment parameters during the historical seeding stage, and the dependent variables are the seeding power and seeding pot position during the historical seeding stage.

[0030] Step 3: Obtain the current segment equipment parameters of the crystal pulling equipment. The current segment equipment parameters include all equipment parameters including batch, segment number, and residual material weight, which are all actual values ​​of the current segment.

[0031] Step 4: Input the current equipment parameters into the optimized model to predict the current seeding power and seeding pot position.

[0032] The automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry according to the embodiment of the present invention includes: a display interface module, a historical equipment parameter storage module, a current section equipment parameter acquisition module, and a prediction module.

[0033] The historical equipment parameter storage module stores historical equipment parameters for the seed crystal rotation and seeding stages, mapped by equipment name. The current-stage equipment parameter acquisition module retrieves the current-stage equipment parameters from the production line. The prediction module inputs the current-stage equipment parameters, invokes the optimized model, and uses model calculations to predict the seeding power and seeding pot position during the seeding stage. The display interface module displays equipment parameter columns and parameter change trends on the system interface.

[0034] Referring to Figures 2 and 3, the equipment parameter fields displayed on the system interface include: equipment name, batch of the local segment, number of segments in the current segment, remaining weight of the current segment, predicted recommended seeding power, and predicted seeding pot position.

[0035] There is a drop-down selection item in the [Device Name] field, which provides the device name for which recommended parameters are required. The user clicks the [Query] button to confirm.

[0036] The [Batch], [Number of Segments], and [Residual Material Weight] fields respectively display the actual values ​​of the equipment's current segment data collection, and the [Seedling Power] and [Seedling Pot Position] fields respectively display the predicted values ​​recommended by the system to the user.

[0037] 2 and 3 , the parameter change trends displayed on the system interface include table form and trend graph form.

[0038] Figure 2 illustrates the parameter change trends in tabular form. The [Batch] column in the table displays the current production batch for the equipment. The [Number of Segments] column displays the current segment and the historical number of segments (for this batch). The [Seedling Power] column displays the predicted value for the current segment, while the historical number of segments displays the actual average power during the seeding phase, for diameters greater than 1000mm. The [Seedling Pot Position] column displays the predicted value for the current segment, while the historical number of segments displays the actual average power during the seeding phase, for diameters greater than 1000mm. The [Residual Weight] column displays the actual residual weight for each segment of the current batch for the equipment.

[0039] Figure 3 shows the parameter change trends in the form of trend charts. Referring to Figures 3 and 4, the seeding power trend chart shows the number of stages on the X-axis, the seeding power on the left, and the residual weight on the right. The three trend lines represent the predicted seeding power, the actual seeding power, and the actual residual weight, respectively. The seeding pot position trend chart shows the number of stages on the X-axis, the seeding pot position on the left, and the residual weight, respectively. The three trend lines represent the predicted seeding pot position, the actual seeding pot position, and the actual residual weight, respectively.

[0040] The above embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for automatically recommending seeding power and seeding pot position in the crystal pulling industry, characterized in that: The following steps are involved: Step 1: Obtain historical equipment parameters of the crystal pulling equipment in the historical seed crystal rotation stage and the crystal seeding stage; Step 2: Input historical equipment parameters into the machine learning model for training and optimization of model performance until the model accuracy reaches more than 95%, and then output the optimized model; Step 3: Obtain the current segment equipment parameters of the crystal pulling equipment; Step 4: Input the current equipment parameters into the optimized model to predict the current seeding power and seeding pot position.

2. The method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to claim 1, characterized in that: in, Among the historical equipment parameters, the equipment parameters of the historical seed crystal rotation stage are all equipment parameters including batch, number of sections, and residual material weight, and the equipment parameters of the historical seed crystal stage are seed crystal power and seed crystal pot position.

3. The method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to claim 2, characterized in that: in, The historical equipment parameters are equipment parameters within one year.

4. The method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to claim 2 or 3, characterized in that: in, The machine learning model adopts the XGboost regression model, and divides the historical equipment parameters into a training set and a validation set. During the model training process, the independent variables are the equipment parameters of the historical seeding stage, and the dependent variables are the seeding power and seeding pot position of the historical seeding stage.

5. The method for automatically recommending seeding power and seeding pot position in the crystal pulling industry according to claim 1, characterized in that: in, The current section equipment parameters of the crystal pulling equipment obtained in step 3 are all equipment parameters including batch, number of sections, and weight of residual material, which are all actual values ​​of the current section.

6. A system for automatically recommending seeding power and seeding pot position in the crystal pulling industry, used to implement the method for automatically recommending seeding power and seeding pot position in the crystal pulling industry as claimed in any one of claims 1 to 5, characterized in that: The system includes: A historical equipment parameter storage module stores historical equipment parameters of the equipment in the seed crystal spinning stage and the seeding stage according to the equipment name; The current segment equipment parameter acquisition module is used to obtain the current segment equipment parameters from the production line; The prediction module inputs the current equipment parameters, calls the optimized model, and predicts the seeding power and seeding pot position in the seeding stage through model calculation; The display interface module is used to display the device parameter fields and parameter change trends in the system interface.

7. The automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry according to claim 6, characterized in that: in, The equipment parameter columns displayed on the system interface include: equipment name, batch of the local segment, number of segments in the current segment, remaining weight of the current segment, predicted recommended seeding power, and predicted seeding pot position.

8. The automatic recommendation system for seeding power and seeding pot position in the crystal pulling industry according to claim 6, characterized in that: in, The parameter change trend includes a table form and a trend graph form.

9. A computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the method for automatically recommending seeding power and seeding pot position in the crystal pulling industry as described in any one of claims 1 to 8.

Images