Wire bow detection method and device, storage medium, electronic equipment and product

Abstract

The invention relates to the technical field of wire bow detection, and particularly provides a wire bow detection method and device, a storage medium, electronic equipment and a product, and the method can comprise the steps that cutting data of silicon rod cutting are obtained, and the cutting data comprise real-time cutting data and cut silicon wafer line data; preprocessing the cut data to obtain processed data; and inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process. According to some embodiments of the invention, the accuracy of wire bow detection can be improved, and the practicability is high.

Description

Technical Field

[0001] The present application relates to the technical field of wire bow detection, and in particular to a method, device, storage medium, electronic device and product for wire bow detection. Background Art

[0002] With the development of modern mechanical processing industry, the requirements for cutting quality and precision are constantly improving, and the requirements for highly intelligent automatic cutting functions are also increasing. In order to ensure the normal operation of the cutting machine and prevent abnormalities, it is usually necessary to detect the wire bow value of the cutting machine to predict risks in time.

[0003] At present, when detecting the wire bow value of the cutting wire during the silicon rod cutting process, the reference point is collected by the monitoring unit, and then the wire bow value is solved by using trigonometric functions in combination with the angle value between the monitoring unit and the wire mesh. However, the range captured by the monitoring unit is limited, and the reference point is not necessarily the lowest point of the wire mesh, so the accuracy of the wire bow value obtained based on this is low.

[0004] Therefore, how to provide a technical solution for an accurate wire bow detection method has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of some embodiments of the present application is to provide a method, device, storage medium, electronic device and product for wire bow detection. Through the technical solutions of the embodiments of the present application, real-time detection of wire bow of a cutting machine can be achieved with high detection efficiency and accuracy.

[0006] In a first aspect, some embodiments of the present application provide a method for wire bow detection, comprising: obtaining cutting data of silicon rod cutting, wherein the cutting data comprises: real-time cutting data and silicon wafer texture data after cutting; preprocessing the cutting data to obtain processed data; inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

[0007] Some embodiments of the present application obtain the cutting data of silicon rod cutting, and input the pre-processed data into the wire bow detection model to obtain the wire bow value corresponding to the real-time cutting process. The embodiments of the present application can realize the real-time detection of wire bow value through model detection, with high detection efficiency and accuracy, and high practicality.

[0008] In some embodiments, the preprocessing of the cutting data to obtain processed data includes: performing a correlation analysis on the cutting data to determine associated feature data; and performing a dimensionality reduction process on the associated feature data to obtain the processed data.

[0009] Some embodiments of the present application perform correlation analysis and dimensionality reduction on cutting data to obtain processed data, thereby improving data support for accurate detection of wire bow values ​​corresponding to real-time cutting processes.

[0010] In some embodiments, before inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to a real-time cutting process, the method further includes: acquiring a training sample data set and a verification sample data set; wherein both the training sample data set and the verification sample data set include: cutting sample data, cutting texture data and actual wire bow values ​​of silicon rod cutting; using the training sample data set to train an initial wire bow model to obtain a wire bow model to be verified; using the verification sample data set to verify the wire bow model to be verified to obtain the wire bow detection model after verification.

[0011] Some embodiments of the present application train and verify the wire bow model through a training sample data set and a verification sample data set to obtain a wire bow detection model, thereby improving model support for accurate detection of wire bow values ​​corresponding to subsequent real-time cutting processes.

[0012] In some embodiments, the obtaining of the training sample data set and the verification sample data set includes: collecting relevant parameters of the historical cutting process of the silicon rod, wherein the relevant parameters include: cutting parameters and texture parameters; analyzing the correlation between the relevant parameters and the actual line bow value, and screening out the associated sample features; performing dimensionality reduction processing on the associated sample features to obtain the cutting sample data and the cutting texture data; and constructing the training sample data set and the verification sample data set based on the cutting sample data, the cutting texture data and the actual line bow value.

[0013] Some embodiments of the present application collect relevant parameters in the historical cutting process of silicon rods, and perform data processing to construct a training sample data set and a verification sample data set, thereby providing effective data support for model training.

[0014] In some embodiments, the cutting parameters include: coolant flow, cutting line parameters, cutting feed speed, return line distance, cutting time, cutting speed and round-trip cycle; the texture parameters include: texture distribution state, silicon wafer thickness and line marks.

[0015] In some embodiments, the use of the training sample data set to train the initial wire bow model to obtain the wire bow model to be verified includes: using the cutting sample data and the cutting texture data in the training sample data set as the input of the initial wire bow model, and the actual wire bow value as the output of the initial wire bow model, training the initial wire bow model to obtain the wire bow model to be verified.

[0016] Some embodiments of the present application train an initial line bow model through a training sample data set to obtain a line bow model to be verified, thereby providing model support for subsequent line bow detection.

[0017] In some embodiments, the wire bow model to be verified is verified using the verification sample data set to obtain the verified wire bow detection model, including: inputting the cutting sample data and the cutting texture data in the verification sample data set into the wire bow model to be verified to obtain a wire bow prediction value; optimizing the wire bow model to be verified based on the error between the wire bow prediction value and the actual wire bow value to determine the verified wire bow detection model.

[0018] Some embodiments of the present application verify the wire bow model to be verified and determine the wire bow detection model to ensure the accuracy of the model and improve the precision of wire bow detection.

[0019] In some embodiments, after inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process, the method further includes: when confirming that the wire bow value exceeds a preset wire bow threshold, adjusting the cutting process during the silicon rod cutting process.

[0020] Some embodiments of the present application compare the wire bow value corresponding to the real-time cutting process with the preset wire bow threshold to determine whether to adjust the cutting process, thereby optimizing the cutting quality and improving production efficiency.

[0021] In a second aspect, some embodiments of the present application provide a device for detecting wire bows, comprising: a data acquisition module for acquiring cutting data of silicon rod cutting, wherein the cutting data comprises: real-time cutting data and silicon wafer texture data after cutting; a data processing module for preprocessing the cutting data to obtain processed data; and a wire bow prediction module for inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

[0022] In a third aspect, some embodiments of the present application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can implement the method described in any embodiment of the first aspect.

[0023] In a fourth aspect, some embodiments of the present application provide an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement a method as described in any embodiment of the first aspect when executing the program.

[0024] In a fifth aspect, some embodiments of the present application provide a computer program product, wherein the computer program product comprises a computer program, wherein the computer program, when executed by a processor, can implement the method described in any embodiment of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of some embodiments of the present application, the drawings required for use in some embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 A schematic diagram of a cutting wire mesh provided for some embodiments of the present application;

[0027] Figure 2 A system diagram of wire bow detection provided for some embodiments of the present application;

[0028] Figure 3 A flow chart of a method for obtaining a wire bow detection model provided in some embodiments of the present application;

[0029] Figure 4 One of the flow charts of the method for wire bow detection provided in some embodiments of the present application;

[0030] Figure 5 A second flowchart of a method for wire bow detection provided in some embodiments of the present application;

[0031] Figure 6 A block diagram of a wire bow detection device provided in some embodiments of the present application;

[0032] Figure 7 A schematic diagram of an electronic device is provided for some embodiments of the present application. DETAILED DESCRIPTION

[0033] The technical solutions in some embodiments of the present application will be described below in conjunction with the drawings in some embodiments of the present application.

[0034] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0035] In related technologies, such as Figure 1 The schematic diagram of the cutting wire network shown in Figure 1The monitoring unit 30 includes a silicon rod 10, a cutting wire mesh 20 (referred to as cutting wire) and a monitoring unit 30. In the process of silicon rod cutting, the monitoring unit 30 can capture the reference points of the silicon rod 10 and the cutting wire mesh 20, and obtain the angle between the monitoring unit 30 and the cutting wire mesh 20. Finally, the reference point and the angle are calculated by the trigonometric function relationship to obtain the wire bow value of the cutting wire mesh 20. The wire bow value can prevent the cutting wire mesh from breaking, jumping and other abnormalities, ensuring that the cutting machine can work normally. However, due to the limited shooting range of the monitoring unit 30, the reference point detected by it is not necessarily the lowest point of the cutting wire mesh 20, and the wire bow value calculated thereby is not very accurate.

[0036] In view of this, some embodiments of the present application provide a method for detecting wire bow, which obtains a wire bow detection model by model training. During the silicon rod cutting process, the cutting data can be obtained, and then the cutting data is pre-processed and input into the wire bow detection model to obtain the real-time wire bow value of the cutting line. Some embodiments of the present application can improve the accuracy of wire bow value detection through the model intelligent detection method, and at the same time improve the detection efficiency, and have high practicality.

[0037] The following is combined with Figure 2 The overall composition structure of the wire bow detection system provided by some embodiments of the present application is exemplified.

[0038] like Figure 2 As shown, some embodiments of the present application provide a system diagram for wire bow detection, and the wire bow detection system may include: a terminal 100 and a detection server 200. Among them, the terminal 100 can obtain the cutting data of the silicon rod in real time, and then the terminal 100 can send the cutting data to the detection server 200. After pre-processing the cutting data, the detection server 200 inputs it into a pre-deployed trained wire bow detection model, and the wire bow detection model can output the wire bow value corresponding to the current real-time cutting process. Through the wire bow value, relevant personnel can timely understand the status of the cutting line to prevent abnormal situations from occurring.

[0039] In some embodiments of the present application, the line bow detection model is pre-deployed in the detection server 200 after the initial line bow model is trained. The terminal 100 can be an image monitoring device or a radar monitoring device for monitoring the cutting line, or can be other types of devices that can collect cutting data during the cutting process, which is not specifically limited in the embodiments of the present application.

[0040] In order to accurately detect the line bow value, it is necessary to first obtain the line bow detection model. Figure 3 The implementation process of obtaining a wire bow detection model provided by some embodiments of the present application is exemplified.

[0041] Please see attached Figure 3 , Figure 3 A flow chart of a method for obtaining a wire bow detection model is provided for some embodiments of the present application. The method for obtaining a wire bow detection model may include:

[0042] S310, obtaining a training sample data set and a verification sample data set; wherein the training sample data set and the verification sample data set both include: cutting sample data, cutting texture data and actual line bow value of silicon rod cutting.

[0043] For example, in some embodiments of the present application, a training sample data set (referred to as training set) and a validation sample data set (referred to as validation set) for model training are first constructed. The training set is used for model training, and the validation set is used for model tuning and performance evaluation.

[0044] In some embodiments of the present application, S310 may include:

[0045] S311, collecting relevant parameters of the historical cutting process of the silicon rod, wherein the relevant parameters include: cutting parameters and grain parameters. The cutting parameters include: coolant flow, cutting line parameters, cutting feed speed, return line distance, cutting time, cutting speed and round trip cycle; the grain parameters include: grain distribution state, silicon wafer thickness and line marks.

[0046] For example, in some embodiments of the present application, relevant parameters in the historical cutting process of the silicon rod are collected. For example, the cutting parameters may include cutting data and cutting processes; wherein the cutting data may include: the flow rate and temperature of the coolant related to the coolant, the steel wire tension and the steel wire diameter related to the steel wire (that is, the cutting wire parameters), etc. The cutting process may include: the feed speed, that is, the moving speed of the silicon rod during the cutting process; the return line distance, that is, the moving distance of the steel wire during the cutting process; the cutting time, that is, the time required for a single cutting; the cutting speed, that is, the speed during the cutting process; the round trip cycle, that is, the cycle of the reciprocating movement of the steel wire, etc. The texture parameters are the texture information of the silicon wafer after cutting, which is obtained by the detector. The texture parameters may include: the texture information obtained by analyzing the silicon wafer texture through image processing technology; the texture distribution state obtained by detecting the uniformity of the texture through transmittance; the silicon wafer thickness obtained by detecting the silicon wafer thickness; the line marks obtained by identifying and measuring the surface of the silicon wafer, etc.

[0047] S312, analyzing the correlation between the relevant parameters and the actual line bow value, and screening out the correlation sample features.

[0048] For example, in some embodiments of the present application, the grey correlation analysis method is used to analyze the above collected related parameters to evaluate the correlation between each parameter feature and the line bow value. Parameter features with a low correlation with the line bow value (for example, the correlation is less than the correlation threshold) are removed to obtain the correlation sample features.

[0049] S313, performing dimensionality reduction processing on the associated sample features to obtain the cutting sample data and the cutting texture data.

[0050] For example, in some embodiments of the present application, the principal component analysis (PCA) method is used to reduce the dimension of the associated sample features to obtain cutting sample data and cutting texture data. PCA can remove noise in the associated sample features, reduce the number of features, and retain the most important components in the data, thereby reducing the complexity of the data and improving the training efficiency of the model.

[0051] S314: construct the training sample data set and the verification sample data set based on the cutting sample data, the cutting texture data and the actual line bow value.

[0052] For example, in some embodiments of the present application, the cutting sample data and the cutting texture data are divided into a training set and a validation set.

[0053] S320, using the training sample data set to train the initial line bow model to obtain the line bow model to be verified.

[0054] For example, in some embodiments of the present application, the XGBoost model (as a specific example of the initial line bow model) is trained using cutting sample data, cutting texture data and actual line bow values ​​to obtain the XGBoost model to be verified. Before training, the model parameters of the XGBoost model are set: the hyperparameters of XGBoost are configured, including the learning rate (learning_rate), the number of trees (n_estimators), the maximum depth (max_depth), the minimum sample weight (min_child_weight), etc. It should be noted that in addition to the XGBoost model, the initial line bow model can also select other types of models, and the embodiments of the present application are not limited to this.

[0055] In some embodiments of the present application, S320 may include: using the cutting sample data and the cutting texture data in the training sample data set as the input of the initial line bow model, and the actual line bow value as the output of the initial line bow model, training the initial line bow model, and obtaining the line bow model to be verified.

[0056] For example, in some embodiments of the present application, the cutting sample data and texture sample data in the training set are used as input, and the actual line bow value is used as output, and training is performed through the fit method of the XGBoost model to obtain the XGBoost model to be verified.

[0057] S330, using the verification sample data set to verify the line bow model to be verified, to obtain the verified line bow detection model.

[0058] For example, in some embodiments of the present application, a validation set is used to evaluate the XGBoost model to be validated to obtain a line bow detection model that meets the requirements.

[0059] In some embodiments of the present application, S330 may include: inputting the cutting sample data and the cutting texture data in the verification sample data set into the line bow model to be verified to obtain a line bow prediction value; optimizing the line bow model to be verified through the error between the line bow prediction value and the actual line bow value, and determining the line bow detection model after verification.

[0060] For example, in some embodiments of the present application, the cutting sample data and the texture sample data in the validation set are used as the input of the XGBoost model to be verified, and the line bow prediction value is output. The mean square error (MSE) or the root mean square error (RMSE) is used to calculate the error between the line bow prediction value and the actual line bow value. According to the error, the model parameters of the XGBoost model to be verified are adjusted to optimize the model performance until a line bow detection model with an error less than the set error threshold is obtained. Among them, the error less than the set error indicates that the model verification has passed.

[0061] The following is combined with Figure 4 The specific process of wire bow detection performed by the detection server 200 provided in some embodiments of the present application is exemplified.

[0062] Please see attached Figure 4 , Figure 4 A flow chart of a method for detecting a wire bow is provided for some embodiments of the present application. The method for detecting a wire bow may include:

[0063] S410, obtaining cutting data of silicon rod cutting, wherein the cutting data includes: real-time cutting data and silicon wafer texture data after cutting.

[0064] For example, in some embodiments of the present application, relevant real-time cutting data and texture data are collected in real time during silicon rod cutting to ensure the accuracy and timeliness of data collection.

[0065] S420, pre-processing the cutting data to obtain processed data.

[0066] In some embodiments of the present application, S420 may include: performing a correlation analysis on the cutting data to determine the correlation feature data; performing a dimensionality reduction process on the correlation feature data to obtain the processed data.

[0067] For example, in some embodiments of the present application, the real-time cutting data and texture data are processed as follows: Figure 3 In the provided method embodiment, the collected cutting parameters and texture parameters are processed by correlation analysis and dimensionality reduction to obtain processed data, thereby ensuring the consistency of data format and model training.

[0068] S430, inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

[0069] For example, in some embodiments of the present application, the processed data is input into a line bow detection model, and the line bow detection model can output a line bow value.

[0070] In some embodiments of the present application, after executing S430, the wire bow detection method may further include: when confirming that the wire bow value exceeds a preset wire bow threshold, adjusting the cutting process during the silicon rod cutting process.

[0071] For example, in some embodiments of the present application, it is determined whether to adjust the cutting process according to the wire bow value obtained by the wire bow detection model. If the wire bow value exceeds the preset wire bow threshold, the cutting process needs to be adjusted. Specifically, the associated parameters of the cutting process (such as coolant flow, feed speed, etc.) can be adjusted to optimize the cutting quality and improve production efficiency.

[0072] The following is combined with Figure 5 The specific process of wire bow detection provided by some embodiments of the present application is exemplified.

[0073] Please see attached Figure 5 , Figure 5 A flow chart of a method for wire bow detection provided for some embodiments of the present application.

[0074] The above process is explained below as an example.

[0075] S510, collecting cutting data, cutting process and texture data of the historical cutting process of the silicon rod.

[0076] S520, analyzing the correlation between the cutting data, cutting process and texture data and the actual wire bow value, and screening out the cutting sample data and cutting texture data.

[0077] S530, constructing a training set and a validation set based on the cutting sample data, the cutting pattern data and the actual line bow value.

[0078] S540, training the initial line bow model using the training set to obtain the line bow model to be verified.

[0079] S550, using the verification set to verify the line bow model to be verified, to obtain a verified line bow detection model.

[0080] S560, obtaining real-time cutting data of silicon rod cutting and texture data of silicon wafers after cutting.

[0081] S570, performing correlation analysis and dimensionality reduction processing on the real-time cutting data and the silicon wafer texture data to obtain processed data.

[0082] S580, input the processed data into a wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

[0083] S590, when the wire bow value exceeds a preset wire bow threshold, adjusting the cutting process during the silicon rod cutting process.

[0084] It can be understood that the specific implementation process of S510 to S590 can refer to the method embodiment provided above, and in order to avoid repetition, the detailed description is appropriately omitted here.

[0085] Through some of the above-mentioned embodiments of the present application, it can be seen that the present application detects the line bow value by means of AI model training, which is not limited by the detection range and other factors, thereby improving the accuracy and sensitivity of the line bow value detection.

[0086] Please refer to Figure 6 , Figure 6 The block diagram of the wire bow detection device provided in some embodiments of the present application is shown. It should be understood that the wire bow detection device corresponds to the above method embodiment and can perform each step involved in the above method embodiment. The specific functions of the wire bow detection device can be found in the above description. To avoid repetition, the detailed description is appropriately omitted here.

[0087] Figure 6 The wire bow detection device includes at least one software function module that can be stored in the memory in the form of software or firmware or solidified in the wire bow detection device, and the wire bow detection device includes: a data acquisition module 610, used to obtain cutting data of silicon rod cutting, wherein the cutting data includes: real-time cutting data and silicon wafer texture data after cutting; a data processing module 620, used to pre-process the cutting data to obtain processed data; a wire bow prediction module 630, used to input the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

[0088] In some embodiments of the present application, the data processing module 620 is used to perform correlation analysis on the cutting data to determine the correlation feature data; and perform dimensionality reduction processing on the correlation feature data to obtain the processed data.

[0089] In some embodiments of the present application, before the data acquisition module 610, the wire bow detection device also includes: a training module (not shown in the figure), which is used to obtain a training sample data set and a verification sample data set; wherein the training sample data set and the verification sample data set both include: cutting sample data, cutting texture data and actual wire bow values ​​of silicon rod cutting; the initial wire bow model is trained using the training sample data set to obtain a wire bow model to be verified; the wire bow model to be verified is verified using the verification sample data set to obtain the wire bow detection model that has passed the verification.

[0090] In some embodiments of the present application, a training module (not shown in the figure) is used to collect relevant parameters of the historical cutting process of silicon rods, wherein the relevant parameters include: cutting parameters and texture parameters; analyzing the correlation between the relevant parameters and the actual line bow value, and screening out related sample features; performing dimensionality reduction processing on the related sample features to obtain the cutting sample data and the cutting texture data; and constructing the training sample data set and the verification sample data set based on the cutting sample data, the cutting texture data and the actual line bow value.

[0091] In some embodiments of the present application, a training module (not shown in the figure) is used to use the cutting sample data and the cutting texture data in the training sample data set as the input of the initial line bow model, and the actual line bow value as the output of the initial line bow model, to train the initial line bow model and obtain the line bow model to be verified.

[0092] In some embodiments of the present application, a training module (not shown in the figure) is used to input the cutting sample data and the cutting texture data in the verification sample data set into the line bow model to be verified to obtain a line bow prediction value; through the error between the line bow prediction value and the actual line bow value, the line bow model to be verified is optimized to determine the line bow detection model after verification.

[0093] In some embodiments of the present application, the wire bow prediction module 630 is used to adjust the cutting process during the silicon rod cutting process when confirming that the wire bow value exceeds a preset wire bow threshold.

[0094] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method, and will not be described in detail here.

[0095] Some embodiments of the present application further provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can implement the operations of the method corresponding to any of the above methods provided in the above embodiments.

[0096] Some embodiments of the present application further provide a computer program product, which includes a computer program, wherein when the computer program is executed by a processor, it can implement the operations corresponding to any of the above methods provided in the above embodiments.

[0097] like Figure 7 As shown, some embodiments of the present application provide an electronic device 700, which includes: a memory 710, a processor 720, and a computer program stored in the memory 710 and executable on the processor 720, wherein the processor 720 can implement a method as described in any of the above embodiments when reading the program from the memory 710 through a bus 730 and executing the program.

[0098] Processor 720 can process digital signals and can include various computing structures, such as complex instruction set computer structure, reduced instruction set computer structure, or a structure that implements a combination of multiple instruction sets. In some examples, processor 720 can be a microprocessor.

[0099] The memory 710 may be used to store instructions executed by the processor 720 or data related to the execution of instructions. These instructions and / or data may include codes for implementing some or all functions of one or more modules described in the embodiments of the present application. The processor 720 of the disclosed embodiment may be used to execute instructions in the memory 710 to implement the method shown above. The memory 710 includes a dynamic random access memory, a static random access memory, a flash memory, an optical memory, or other memory known to those skilled in the art.

[0100] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0101] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0102] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

Claims

1. A method for detecting a wire bow, characterized in that: include: Acquire cutting data of silicon rod cutting, wherein the cutting data includes: real-time cutting data and silicon wafer texture data after cutting; Preprocessing the cutting data to obtain processed data; The processed data is input into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

2. The method according to claim 1, characterized in that The preprocessing of the cutting data to obtain processed data includes: Performing correlation analysis on the cutting data to determine correlation feature data; Performing dimension reduction processing on the associated feature data to obtain the processed data.

3. The method according to claim 1 or 2, characterized in that Before inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process, the method further includes: Obtaining a training sample data set and a verification sample data set; wherein the training sample data set and the verification sample data set both include: cutting sample data, cutting texture data and actual line bow value of silicon rod cutting; Using the training sample data set to train the initial line bow model to obtain the line bow model to be verified; The line bow model to be verified is verified using the verification sample data set to obtain the verified line bow detection model.

4. The method according to claim 3, characterized in that The step of obtaining a training sample data set and a verification sample data set includes: Collect relevant parameters of the historical cutting process of the silicon rod, wherein the relevant parameters include: cutting parameters and texture parameters; Analyze the correlation between the relevant parameters and the actual line bow value, and screen out the related sample features; Performing dimensionality reduction processing on the associated sample features to obtain the cutting sample data and the cutting pattern data; The training sample data set and the verification sample data set are constructed based on the cutting sample data, the cutting texture data and the actual line bow value.

5. The method according to claim 4, characterized in that The cutting parameters include: coolant flow, cutting line parameters, cutting feed speed, return line distance, cutting time, cutting speed and round trip cycle; the texture parameters include: texture distribution state, silicon wafer thickness and line marks.

6. The method according to claim 4 or 5, characterized in that The using the training sample data set to train the initial line bow model to obtain the line bow model to be verified includes: The cutting sample data and the cutting texture data in the training sample data set are used as inputs of the initial line bow model, and the actual line bow value is used as output of the initial line bow model. The initial line bow model is trained to obtain the line bow model to be verified.

7. The method according to claim 4 or 5, characterized in that The method of using the verification sample data set to verify the line bow model to be verified to obtain the verified line bow detection model includes: Inputting the cutting sample data and the cutting pattern data in the verification sample data set into the line bow model to be verified to obtain a line bow prediction value; The line bow model to be verified is optimized by the error between the line bow prediction value and the actual line bow value, and the line bow detection model after verification is determined.

8. The method according to any one of claims 1-2, 4-5, characterized in that: After inputting the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process, the method further includes: When it is confirmed that the line bow value exceeds a preset line bow threshold, the cutting process in the silicon rod cutting process is adjusted.

9. A wire bow detection device, characterized in that: include: A data acquisition module, used to acquire cutting data of silicon rod cutting, wherein the cutting data includes: real-time cutting data and silicon wafer texture data after cutting; A data processing module, used for preprocessing the cutting data to obtain processed data; The wire bow prediction module is used to input the processed data into a pre-trained wire bow detection model to obtain a wire bow value corresponding to the real-time cutting process.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program executes the method according to any one of claims 1 to 8 when executed by a processor.

11. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the computer program executes the method according to any one of claims 1 to 8 when being run by the processor.

12. A computer program product, characterized in that The computer program product comprises a computer program, wherein the computer program executes the method according to any one of claims 1 to 8 when executed by a processor.

Images