Method and system for predicting wet process parameters for solar cell manufacturing

Abstract

The application provides a wet process parameter prediction method and system for solar cell manufacturing, and belongs to the technical field of parameter optimization. The wet process parameter prediction method for solar cell manufacturing comprises the following steps: dividing a wet process into a pretreatment process, a purification process, an etching process and a post-treatment process according to the processes of the wet process; respectively creating a pretreatment prediction model, a purification prediction model, an etching prediction model and a post-treatment prediction model, and obtaining a fitness function; obtaining each preset process parameter in the wet process; the process parameters comprise a silicon wafer type, purification liquid parameters, etching liquid parameters and cleaning liquid parameters; inputting the fitness function and each preset process parameter into the process parameter prediction model, obtaining an initial population according to the preset process parameters, and guiding the evolution of the initial population according to the fitness function to search for and obtain optimized process parameters. The stability and consistency of the wet process in the solar cell manufacturing process are significantly improved.

Description

Technical Field

[0001] This application relates to the field of parameter optimization technology, specifically to a method and system for predicting wet process parameters in solar cell manufacturing. Background Technology

[0002] In the field of solar cell manufacturing, wet processing involves several key steps, including cleaning, purification, etching, and subsequent chemical treatment of silicon wafers. Among these, wet etching uses a specific chemical solution to react with the silicon wafer surface, thereby precisely removing excess material and forming the desired circuit patterns and structures.

[0003] However, even slight fluctuations in process parameters during wet processing can significantly impact the performance of the final product. Therefore, accurately predicting and controlling wet process parameters to ensure the consistency and stability of solar cell manufacturing has become a pressing issue. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a method and system for predicting wet process parameters in solar cell manufacturing.

[0005] In a first aspect, this application provides a method for predicting wet process parameters in solar cell manufacturing, including:

[0006] Based on the process flow of wet processes, wet processes are divided into pretreatment processes, purification processes, etching processes, and post-treatment processes.

[0007] A pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-processing prediction model corresponding to the post-processing process are created respectively, and an appropriate value function is obtained based on the pretreatment prediction model, the purification prediction model, the etching prediction model, and the post-processing prediction model.

[0008] Obtain the preset process parameters for the wet process; the process parameters include silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters;

[0009] The optimal process parameters are obtained by inputting the optimality function and each process parameter into the process parameter prediction model, obtaining an initial population based on the process parameters, and guiding the evolution of the initial population according to the optimality function.

[0010] Optionally, obtaining the initial population based on the process parameters includes:

[0011] The limitation ranges of the purification solution parameters, the etching solution parameters, and the cleaning solution parameters are obtained respectively;

[0012] Obtain the number of individuals in the initial population;

[0013] Based on the number of individuals, initial individuals of the purification solution parameters, the etching solution parameters, and the cleaning solution parameters are generated at equal intervals within the restricted range to obtain an initial population.

[0014] Optionally, the step of guiding the initial population evolution according to the fitness value function to search for optimized process parameters includes:

[0015] The initial population is input into the fitness function to output the predicted quality parameters of the finished product;

[0016] The predicted quality parameters of the finished product guide the evolution of the initial population to search for optimized process parameters.

[0017] Optionally, inputting the initial population into the fitness function to output the predicted quality parameters of the finished product includes:

[0018] Each initial individual in the initial population is input into the fitness function; the initial individual includes several purification solution parameters, several etching solution parameters, and several cleaning solution parameters.

[0019] The silicon wafer type and the cleaning fluid parameters are input into the pretreatment prediction model to output the cleanliness prediction data of the silicon wafer surface and the cleaning fluid residue prediction data.

[0020] The purification solution parameters, silicon wafer surface cleanliness prediction data, and cleaning solution residue prediction data are input into the purification prediction model to output the silicon wafer predicted purity.

[0021] The predicted purity of the silicon wafer and the parameters of the etching solution are input into the etching prediction model to output the predicted etching depth and the predicted etching residue data.

[0022] The cleaning fluid parameters, the etching prediction depth, and the etching residue prediction data are input into the post-processing prediction model to output the predicted cleanliness and predicted flatness of the silicon wafer product.

[0023] The predicted quality parameters of the finished product are obtained based on the predicted etching depth, the predicted cleanliness of the silicon wafer, and the predicted flatness of the silicon wafer.

[0024] Optionally, obtaining the predicted quality parameters of the finished product based on the predicted etching depth, the predicted cleanliness of the silicon wafer, and the predicted flatness of the silicon wafer includes:

[0025] The first weight for predicting etching depth, the second weight for predicting silicon wafer cleanliness, and the third weight for predicting silicon wafer flatness are obtained.

[0026] The predicted etching depth, predicted cleanliness of the silicon wafer, and predicted flatness of the silicon wafer are weighted and summed according to the first weight, the second weight, and the third weight to obtain the appropriate value.

[0027] Optionally, the step of guiding the evolution of the initial population based on the predicted quality parameters of the finished product to search for optimized process parameters includes:

[0028] From the initial population, the initial sample with the largest predicted quality parameter of the finished product is determined as the target sample;

[0029] The target sample is subjected to crossover and mutation operations to obtain an evolved sample, and the evolved sample is input into the fitness function until the predicted quality parameters of the evolved sample reach the preset conditions to obtain an optimized population.

[0030] The optimized process parameters are determined based on the predicted quality parameters of the finished products for each individual in the optimized population.

[0031] Optionally, the step of creating a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-processing prediction model corresponding to the post-processing process includes:

[0032] Construct a training dataset; wherein the training dataset includes process parameters and corresponding measurement data of silicon wafer surface cleanliness, cleaning fluid residue, silicon wafer purity, etching depth, etching residue, silicon wafer product cleanliness, and silicon wafer product flatness.

[0033] The first preset neural network model is trained based on the training dataset to obtain the preprocessed prediction model;

[0034] The second preset neural network model is trained based on the training dataset to obtain the purified prediction model;

[0035] The etching prediction model is obtained by training the third preset neural network model based on the training dataset.

[0036] The fourth preset neural network model is trained based on the training dataset to obtain the post-processing prediction model.

[0037] Optionally, training the first preset neural network model based on the training dataset to obtain the preprocessed prediction model includes:

[0038] The cleanliness measurement data and the cleaning fluid residue prediction parameters in the training dataset are used as the first training labels, and the silicon wafer type and the cleaning fluid parameters in the process parameters are used as the first training samples.

[0039] The first training sample is input into the first preset neural network model to obtain the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters;

[0040] Based on the first training label, the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters, the first preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the preprocessed prediction model.

[0041] The step of training the second preset neural network model based on the training dataset to obtain the refined prediction model includes:

[0042] The silicon wafer purity measured in the training dataset is used as the second training label, and the cleaning fluid parameter prediction data, the cleaning fluid parameter prediction parameters, and the purification fluid parameters in the process parameters are used as the second training samples.

[0043] The second training sample is input into the second preset neural network model to obtain the predicted purity of the silicon wafer;

[0044] Based on the second training label and the predicted purity of the silicon wafer, the second preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the purification prediction model.

[0045] The step of training the third preset neural network model based on the training dataset to obtain the etching prediction model includes:

[0046] The cleanliness measurement data in the training dataset is used as the third training label, and the predicted purity of the silicon wafer and the etching solution parameters in the process parameters are used as the third training samples.

[0047] The third training sample is input into the third preset neural network model to obtain the etching prediction depth and etching residue prediction data;

[0048] Based on the third training label, the etching prediction depth and the etching residue prediction data, the third preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the etching prediction model.

[0049] The step of training the fourth preset neural network model based on the training dataset to obtain the post-processing prediction model includes:

[0050] The silicon wafer product cleanliness and silicon wafer product flatness in the training dataset are used as the fourth training labels, and the etching prediction depth, etching residue prediction data and cleaning fluid parameters in the process parameters are used as the fourth training samples.

[0051] The fourth training sample is input into the fourth preset neural network model to obtain the predicted cleanliness and predicted flatness of the silicon wafer product.

[0052] Based on the fourth training label, the predicted cleanliness of the silicon wafer product, and the predicted flatness of the silicon wafer product, the fourth preset neural network model is iteratively updated until the iteration termination condition is reached, thus obtaining the post-processing prediction model.

[0053] Secondly, in one embodiment, this application provides a wet process parameter prediction system for solar cell manufacturing, comprising:

[0054] The process division module is used to divide the wet process into pretreatment process, purification process, etching process and post-treatment process according to the process of the wet process.

[0055] The function generation module is used to create a preprocessing prediction model corresponding to the preprocessing process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a postprocessing prediction model corresponding to the postprocessing process, and to obtain an appropriate value function based on the preprocessing prediction model, the purification prediction model, the etching prediction model, and the postprocessing prediction model.

[0056] The process parameter acquisition module is used to acquire various preset process parameters in the wet process; the process parameters include silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters.

[0057] The prediction parameter generation module is used to input the fitness value function and each of the process parameters into the process parameter prediction model, obtain an initial population based on the process parameters, and guide the evolution of the initial population based on the fitness value function to search for and obtain optimized process parameters.

[0058] Thirdly, in one embodiment, this application provides an electronic device including a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to perform the steps in the wet process parameter prediction method for solar cell manufacturing in any of the above embodiments.

[0059] Fourthly, in one embodiment, this application provides a storage medium storing a computer program that is loaded by a processor to execute the steps in the wet process parameter prediction method for solar cell manufacturing in any of the above embodiments.

[0060] In summary, this application significantly improves the prediction accuracy of each stage of the wet process by finely dividing the wet process into pretreatment, purification, etching, and post-treatment processes, and creating corresponding prediction models for each process (pretreatment prediction model, purification prediction model, etching prediction model, and post-treatment prediction model). The appropriate value functions obtained from each prediction model accurately reflect the complex relationship between process parameters and output results. Furthermore, by acquiring key process parameters in the wet process, including silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters, and inputting these parameters along with the appropriate value functions into the process parameter prediction model, the initial population is iteratively evolved using the appropriate value functions. This continuous search and optimization of process parameters improves the search efficiency and determines the optimal combination of process parameters, thereby significantly enhancing the stability and consistency of the wet process in solar cell manufacturing. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of an application scenario in one embodiment of this application.

[0063] Figure 2 This is a flowchart of a wet process parameter prediction method for solar cell manufacturing in one embodiment of this application;

[0064] Figure 3 This is a flowchart of a method for generating predicted quality parameters of finished products in one embodiment of this application;

[0065] Figure 4 This is a flowchart of a method for determining optimized process parameters in one embodiment of this application;

[0066] Figure 5 This is a flowchart of a training method in one embodiment of this application;

[0067] Figure 6 This is a block diagram of a wet process parameter prediction system for solar cell manufacturing in one embodiment of this application. Detailed Implementation

[0068] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0069] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified. In this application, the term "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use this application. In the following description, details are set forth for illustrative purposes. It should be understood that those skilled in the art will recognize that this application can be implemented without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessary detail that would obscure the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0070] The wet process parameter prediction method for solar cell manufacturing in this application embodiment is applied to the wet process parameter prediction system for solar cell manufacturing. The wet process parameter prediction system for solar cell manufacturing is set in an electronic device. The electronic device can be a terminal, such as a mobile phone or a tablet computer. The electronic device can also be a server or a service cluster composed of multiple servers.

[0071] like Figure 1 As shown, Figure 1 This is a schematic diagram illustrating an application scenario of the wet process parameter prediction method for solar cell manufacturing in this application embodiment. The application scenario of the wet process parameter prediction method for solar cell manufacturing in this application embodiment includes an electronic device 100, which integrates a wet process parameter prediction system for solar cell manufacturing. The electronic device 100 runs a computer-readable storage medium corresponding to the wet process parameter prediction method for solar cell manufacturing to execute the steps of the wet process parameter prediction method for solar cell manufacturing.

[0072] Understandable, Figure 1The electronic devices in the application scenario of the wet process parameter prediction method for solar cell manufacturing, or the devices contained in the electronic devices, do not constitute a limitation on the embodiments of this application. That is, the number or type of equipment in the application scenario of the wet process parameter prediction method for solar cell manufacturing, or the number or type of devices contained in each equipment, do not affect the overall implementation of the technical solution in the embodiments of this application, and can all be considered as equivalent substitutions or derivatives of the technical solutions claimed in the embodiments of this application.

[0073] In this application embodiment, the electronic device 100 can be an independent device, or a device network or device cluster composed of devices. For example, the electronic device 100 described in this application embodiment includes, but is not limited to, a computer, a network host, a single network device, a set of multiple network devices, or a cloud device composed of multiple devices. Among them, the cloud device is composed of a large number of computers or network devices based on cloud computing.

[0074] Those skilled in the art will understand that Figure 1 The application scenarios shown are merely one application scenario corresponding to the technical solution of this application, and do not constitute a limitation on the application scenarios of the technical solution of this application. Other application scenarios may include more than one application scenario. Figure 1 The number of more or fewer electronic devices shown, or the network connections of electronic devices, for example Figure 1 Only one electronic device is shown in the diagram. It is understood that the scenario of the wet process parameter prediction method for solar cell manufacturing may also include one or more other electronic devices, which are not specifically limited here. The electronic device 100 may also include a memory for storing information related to the wet process parameter prediction method for solar cell manufacturing.

[0075] Furthermore, in the application scenario of the wet process parameter prediction method for solar cell manufacturing in this application embodiment, the electronic device 100 can be equipped with a display device, or the electronic device 100 can be without a display device but connected to an external display device 200. The display device 200 is used to output the results of the execution of the wet process parameter prediction method for solar cell manufacturing in the electronic device. The electronic device 100 can access the background database 300, which can be the local storage of the electronic device 100 or it can be located in the cloud. The background database 300 stores information related to the wet process parameter prediction method for solar cell manufacturing.

[0076] It should be noted that, Figure 1The application scenario of the wet process parameter prediction method for solar cell manufacturing shown is merely an example. The application scenario of the wet process parameter prediction method for solar cell manufacturing described in this application embodiment is to more clearly illustrate the technical solution of this application embodiment and does not constitute a limitation on the technical solution provided in this application embodiment.

[0077] Based on the application scenarios of the above-mentioned method for predicting wet process parameters in solar cell manufacturing, an embodiment of the method for predicting wet process parameters in solar cell manufacturing is proposed.

[0078] Firstly, such as Figure 2 As shown, in one embodiment, this application provides a method for predicting wet process parameters in solar cell manufacturing. The method for predicting wet process parameters in solar cell manufacturing includes:

[0079] Step S101: Based on the wet process, the wet process is divided into pretreatment process, purification process, etching process and post-treatment process.

[0080] As an example, the main purpose of the pretreatment process is to remove stains and impurities from the silicon wafer surface, providing a clean substrate for subsequent purification and etching stages. A cleaning solution is used to chemically clean the silicon wafer to remove contaminants such as grease, dust, and metal ions. The purification process aims to further improve the purity of the silicon wafer, removing trace impurities that may remain from the pretreatment process. The etching process involves etching the silicon wafer surface according to a predetermined pattern to form the desired circuit structure, achieved through chemical or physical etching. The post-treatment process primarily removes residues generated during etching and performs final cleaning and drying of the silicon wafer to ensure the cleanliness and flatness of the finished product.

[0081] Step S102: Create a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-treatment prediction model corresponding to the post-treatment process, and obtain the appropriate value function based on the pretreatment prediction model, purification prediction model, etching prediction model, and post-treatment prediction model.

[0082] Step S103: Obtain the preset process parameters for each step of the wet process; the process parameters include silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters;

[0083] Silicon wafer types can include monocrystalline silicon wafers, polycrystalline silicon wafers, thin-film silicon wafers, and special types of silicon wafers. Monocrystalline silicon wafers are made of high-purity monocrystalline silicon and have high crystal integrity and uniformity. Polycrystalline silicon wafers are composed of multiple small silicon crystal particles, and their crystal integrity and uniformity are slightly inferior to those of monocrystalline silicon wafers. However, polycrystalline silicon wafers have relatively low production costs, so they are widely used in solar cell manufacturing. Thin-film silicon wafers are thinner silicon wafers, usually prepared on a substrate material by methods such as chemical vapor deposition (CVD) or physical vapor deposition (PVD). Special types of silicon wafers include silicon-based heterojunction (HJT) silicon wafers and silicon wafers for perovskite / silicon tandem cells.

[0084] The cleaning fluid parameters may include the type, concentration, cleaning temperature, and cleaning time; the purification fluid parameters may include the type, initial concentration, purification temperature, and purification time; and the etching fluid parameters may include the type, initial concentration, etching temperature, and etching time.

[0085] Step S104: Input the fitness value function and each preset process parameter into the process parameter prediction model, obtain the initial population according to the preset process parameters, and guide the evolution of the initial population according to the fitness value function to search for the optimal process parameters.

[0086] In the above embodiments, by finely dividing the wet process into pretreatment, purification, etching, and post-treatment processes, and creating corresponding prediction models for each process (pretreatment prediction model, purification prediction model, etching prediction model, and post-treatment prediction model), the prediction accuracy of each stage of the wet process is significantly improved. The appropriate value functions obtained from each prediction model accurately reflect the complex relationship between process parameters and output results. Furthermore, by acquiring key process parameters in the wet process, including silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters, and inputting these parameters along with the appropriate value functions into the process parameter prediction model, the initial population is iteratively evolved using the appropriate value functions. This continuously searches for and optimizes process parameters, improving the search efficiency and determining the optimal combination of process parameters, thereby significantly improving the stability and consistency of the wet process in solar cell manufacturing.

[0087] In some embodiments, obtaining an initial population based on preset process parameters includes:

[0088] Step S201: Obtain the limit ranges for the purification solution parameters, etching solution parameters, and cleaning solution parameters respectively;

[0089] Step S202: Obtain the number of individuals in the initial population;

[0090] Step S203: Based on the number of individuals, generate initial individuals with purification solution parameters, etching solution parameters, and cleaning solution parameters at equal intervals within the restricted range to obtain the initial population.

[0091] In the above implementation, firstly, the constraint ranges for the purification solution parameters, etching solution parameters, and cleaning solution parameters were obtained, ensuring that the subsequently generated process parameters were all within a reasonable and safe range, effectively avoiding excessive deviations in the final predicted parameters due to parameters exceeding the constraints. Secondly, by determining the number of individuals in the initial population, a reasonable number of individuals not only ensures the comprehensiveness of the search but also avoids over-computation and resource waste while ensuring search efficiency. Finally, based on the determined number of individuals, initial individuals for the purification solution parameters, etching solution parameters, and cleaning solution parameters were generated at equal intervals within the constraint range, thereby constructing a uniformly distributed initial population to ensure broad coverage of the population in the search space and increase the probability of finding the global optimum.

[0092] In some embodiments, the initial population evolution is guided by a fitness value function to search for optimal process parameters, including:

[0093] Step S301: Input the initial population into the fitness function to output the predicted quality parameters of the finished product;

[0094] Step S302: Based on the predicted quality parameters of the finished product, guide the evolution of the initial population to search for and obtain the optimized process parameters.

[0095] Reference Figure 3 In some embodiments, step S301 includes:

[0096] Step S3011: Input each initial individual in the initial population into the fitness function; the initial individual includes several purification solution parameters, several etching solution parameters, and several cleaning solution parameters;

[0097] Step S3012: Input the silicon wafer type and cleaning solution parameters into the pretreatment prediction model to output the cleanliness prediction data of the silicon wafer surface and the cleaning solution residue prediction data.

[0098] Step S3013: Input the purification solution parameters, the predicted cleanliness data of the silicon wafer surface, and the predicted cleaning solution residue data into the purification prediction model to output the predicted purity of the silicon wafer.

[0099] Step S3014: Input the predicted purity of the silicon wafer and the etching solution parameters into the etching prediction model to output the predicted etching depth and etching residue prediction data;

[0100] Step S3015: Input the cleaning fluid parameters, etching prediction depth and etching residue prediction data into the post-processing prediction model to output the predicted cleanliness and predicted flatness of the silicon wafer product.

[0101] Step S3016: Obtain the predicted quality parameters of the finished product based on the predicted etching depth, predicted cleanliness of the silicon wafer, and predicted flatness of the silicon wafer.

[0102] In the above implementation, firstly, each initial individual in the initial population (including purification solution parameters, etching solution parameters, and cleaning solution parameters) is input into a suitability function to achieve a preliminary evaluation of the process parameters, thereby screening out potential combinations of working parameters. Next, the pretreatment prediction model predicts the cleanliness of the silicon wafer surface and the amount of cleaning solution residue based on the silicon wafer type and cleaning solution parameters. Simultaneously, the output of the pretreatment prediction model, namely the predicted cleanliness data of the silicon wafer surface and the predicted cleaning solution residue data, serves as the input to the purification prediction model. The purification prediction model uses the purification solution parameters, the predicted cleanliness data of the silicon wafer surface, and the predicted cleaning solution residue data to output the predicted purity of the silicon wafer, reflecting not only the effectiveness of the purification process but also serving as one of the important parameters input to the subsequent etching prediction model. The etching prediction model predicts the etching depth and the amount of etching residue based on the predicted purity of the silicon wafer and the etching solution parameters. The post-processing prediction model outputs predicted cleanliness and flatness of the silicon wafer product based on cleaning solution parameters, predicted etching depth, and predicted etching residue data. In the post-processing prediction model, the predicted etching depth and etching residue data are derived from the output of the etching prediction model. This process fully considers the correlation between the inputs of the pre-processing prediction model, purification prediction model, etching prediction model, and post-processing prediction model. Finally, the predicted etching depth, predicted cleanliness of the silicon wafer product, and predicted flatness of the silicon wafer product are combined to obtain the predicted quality parameters of the finished product. These predicted quality parameters are derived from the outputs of each model (pre-processing prediction model, purification prediction model, etching prediction model, and post-processing prediction model), fully reflecting the continuity and cumulative effect of each model. This allows the predicted quality parameters to provide accurate feedback for the optimization of the entire wet process; therefore, the predicted quality parameters are used as the appropriate value output by the appropriate value function.

[0103] In some embodiments, step S3016 includes:

[0104] Step S30161: Obtain the first weight of the etching prediction depth, the second weight of the silicon wafer product prediction cleanliness, and the third weight of the silicon wafer product prediction flatness;

[0105] Step S30162: Based on the first weight, the second weight, and the third weight, the predicted etching depth, the predicted cleanliness of the silicon wafer product, and the predicted flatness of the silicon wafer product are weighted and summed to obtain an appropriate value.

[0106] Reference Figure 4 In some embodiments, the initial population evolution is guided by the predicted quality parameters of the finished product to search for optimized process parameters, including:

[0107] Step S401: Identify the initial sample with the largest predicted quality parameter of the finished product from the initial population as the target sample;

[0108] Step S402: Perform crossover and mutation operations on the target sample to obtain the evolved sample, and input the evolved sample into the fitness function until the predicted quality parameters of the evolved sample reach the preset conditions to obtain the optimized population;

[0109] It's important to note that each cross-mutation operation on the target sample only applies to one specific process parameter. The cross-mutation parameter can be input into the corresponding pretreatment prediction model, purification prediction model, etching prediction model, and post-treatment prediction model based on the process in which it occurs. Then, the model corresponding to the process following that step is executed. For example, if the cross-mutation parameter is in the purification process, it is input into the purification prediction model, and then the etching and post-treatment prediction models are executed sequentially without recalculating the output of the pretreatment prediction model (since its output remains unchanged). Similarly, if the cross-mutation parameter is in the etching process, it is input into the etching prediction model, and then the post-treatment prediction model is executed without recalculating the outputs of the pretreatment and purification prediction models. Likewise, if the cross-mutation parameter is in the post-treatment process, it is input only into the post-treatment prediction model without recalculating the outputs of the pretreatment, purification, and etching prediction models.

[0110] Step S403: Determine the optimized process parameters based on the predicted quality parameters of the finished products for each individual in the optimized population.

[0111] In the above implementation, firstly, the evolutionary process of the initial population is guided by the predicted quality parameters of the finished product. Initial samples with higher predicted quality parameters are selected from the initial population as target samples, ensuring that subsequent evolution is based on samples with high-quality parameters, thereby improving optimization efficiency. Secondly, the target samples undergo crossover and mutation operations to obtain evolved samples. These evolved samples are then input into a fitness function for further evaluation. By simulating the mechanisms of natural selection and genetic variation, the evolved samples inherit the superior genes of the target samples while introducing new mutations to increase the diversity of the search space. Through continuous iteration, the predicted quality parameters of the evolved samples reach preset conditions, ultimately resulting in an optimized population. Finally, the optimized process parameters are determined based on the predicted quality parameters of each individual in the optimized population.

[0112] Reference Figure 5 In some embodiments, step S102 includes:

[0113] Step S1021: Construct the training dataset;

[0114] The training dataset includes process parameters and corresponding measurements of silicon wafer surface cleanliness, cleaning fluid residue, silicon wafer purity, etching depth, etching residue, silicon wafer product cleanliness, and silicon wafer product flatness.

[0115] Step S1022: Train the first preset neural network model based on the training dataset to obtain a preprocessed prediction model;

[0116] Step S1023: Train the second preset neural network model based on the training dataset to obtain a purified prediction model;

[0117] Step S1024: Train the third preset neural network model based on the training dataset to obtain the etching prediction model;

[0118] Step S1025: Train the fourth preset neural network model based on the training dataset to obtain the post-processing prediction model.

[0119] In some embodiments, step S1022 includes:

[0120] Step S10221: Use the cleanliness measurement data and cleaning fluid residue prediction parameters in the training dataset as the first training labels, and use the silicon wafer type and cleaning fluid parameters in the process parameters as the first training samples;

[0121] Step S10222: Input the first training sample into the first preset neural network model to obtain the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters;

[0122] Step S10223: Based on the first training label, the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters, iteratively update the first preset neural network model until the iteration termination condition is reached to obtain the preprocessed prediction model;

[0123] In some embodiments, step S1023 includes:

[0124] Step S10231: Use the silicon wafer purity measurement in the training dataset as the second training label, and use the cleaning solution parameter prediction data, the cleaning solution parameter prediction parameters, and the purification solution parameters in the process parameters as the second training samples;

[0125] Step S10232: Input the second training sample into the second preset neural network model to obtain the predicted purity of the silicon wafer;

[0126] Step S10233: Based on the second training label and the silicon wafer predicted purity, iteratively update the second preset neural network model until the iteration termination condition is reached to obtain the purification prediction model.

[0127] In some embodiments, step S1024 includes:

[0128] Step S10241: Use the cleanliness measurement data in the training dataset as the third training label, and use the silicon wafer predicted purity and the etching solution parameters in the process parameters as the third training samples.

[0129] Step S10242: Input the third training sample into the third preset neural network model to obtain the etching prediction depth and etching residue prediction data;

[0130] Step S10243: Based on the third training label, etching prediction depth and etching residue prediction data, iteratively update the third preset neural network model until the iteration termination condition is reached to obtain the etching prediction model.

[0131] In some embodiments, step S1025 includes:

[0132] Step S10251: Use the measured cleanliness and flatness of silicon wafer products in the training dataset as the fourth training labels, and use the etching prediction depth, etching residue prediction data and cleaning fluid parameters in the process parameters as the fourth training samples.

[0133] Step S10252: Input the fourth training sample into the fourth preset neural network model to obtain the predicted cleanliness and predicted flatness of the silicon wafer product;

[0134] Step S10253: Based on the fourth training label, the predicted cleanliness of the silicon wafer product, and the predicted flatness of the silicon wafer product, the fourth preset neural network model is iteratively updated until the iteration termination condition is reached, and the post-processing prediction model is obtained.

[0135] Thus, the first, second, third, and fourth preset neural network models can all be trained independently to obtain the preprocessing prediction model, purification prediction model, etching prediction model, and post-processing prediction model. When the processes in the wet process change, only the preset neural network model for the corresponding process needs to be trained separately, and the other models can be reused.

[0136] Reference Figure 6 Secondly, in one embodiment, this application provides a wet process parameter prediction system for solar cell manufacturing, comprising:

[0137] The process division module is used to divide the wet process into pretreatment process, purification process, etching process and post-treatment process according to the process of the wet process.

[0138] The function generation module is used to create a preprocessing prediction model corresponding to the preprocessing process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a postprocessing prediction model corresponding to the postprocessing process, and obtain the appropriate value function based on the preprocessing prediction model, purification prediction model, etching prediction model, and postprocessing prediction model.

[0139] The process parameter acquisition module is used to acquire the preset process parameters in the wet process; the process parameters include silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters.

[0140] The prediction parameter generation module is used to input the moderation function and each preset process parameter into the process parameter prediction model, obtain the initial population based on the preset process parameters, and guide the evolution of the initial population according to the moderation function to search for the optimal process parameters.

[0141] Thirdly, in one embodiment, this application provides an electronic device illustrating the structure of the electronic device involved in this application, specifically:

[0142] The electronic device may include components such as a processor 401 with one or more processing cores, a memory 402 with one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will understand that... Figure 5 The structure of the electronic device shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0143] The processor 401 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 402, and by calling data stored in the memory 402, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Optionally, the processor 401 may include one or more processing cores; preferably, the processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and computer programs, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 401.

[0144] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, computer programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the server, etc. In addition, the memory 402 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0145] The electronic device also includes a power supply 403 that supplies power to the various components. Preferably, the power supply 403 can be logically connected to the processor 401 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 403 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0146] The electronic device may also include an input unit 404, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0147] Although not shown, the electronic device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, when the electronic device is a model training electronic device, the processor 401 in the electronic device will load the executable files corresponding to the processes of one or more computer programs into the memory 402 according to the following instructions, and the processor 401 will run the computer programs stored in the memory 402 to perform the above steps.

[0148] Those skilled in the art will understand that all or part of the steps in any of the methods in the above embodiments can be performed by a computer program or by a computer program controlling related hardware. The computer program can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0149] Fourthly, in one embodiment, this application provides a storage medium storing a plurality of computer programs that can be loaded by a processor to perform the above steps.

[0150] It will be understood by those skilled in the art that any references to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchlink, DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0151] Since the computer program stored in the storage medium can execute the steps in the wet process parameter prediction method for solar cell manufacturing in any embodiment of the present application, the beneficial effects that the wet process parameter prediction method for solar cell manufacturing in any embodiment of the present application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0152] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0153] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed descriptions of other embodiments above, which will not be repeated here.

[0154] The above provides a detailed description of the wet process parameter prediction method and system for solar cell manufacturing provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. 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 this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0155] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A method for predicting wet process parameters in solar cell manufacturing, characterized in that, include: Based on the process flow of wet processes, wet processes are divided into pretreatment processes, purification processes, etching processes, and post-treatment processes. Create a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-processing prediction model corresponding to the post-processing process, respectively. Obtain the preset process parameters for the wet process; The process parameters include silicon wafer type, purification solution parameters, etching solution parameters, and cleaning solution parameters. An initial population is obtained based on the process parameters. The silicon wafer type and the cleaning fluid parameters are input into the pretreatment prediction model to output the cleanliness prediction data of the silicon wafer surface and the cleaning fluid residue prediction data. The purification solution parameters, silicon wafer surface cleanliness prediction data, and cleaning solution residue prediction data are input into the purification prediction model to output the silicon wafer predicted purity. The predicted purity of the silicon wafer and the parameters of the etching solution are input into the etching prediction model to output the predicted etching depth and the predicted etching residue data. The cleaning fluid parameters, the etching prediction depth, and the etching residue prediction data are input into the post-processing prediction model to output the predicted cleanliness and predicted flatness of the silicon wafer product. The first weight for predicting etching depth, the second weight for predicting silicon wafer cleanliness, and the third weight for predicting silicon wafer flatness are obtained. Based on the first weight, the second weight, and the third weight, the predicted etching depth, the predicted cleanliness of the silicon wafer product, and the predicted flatness of the silicon wafer product are weighted and summed to obtain an appropriate value; The initial population is evolved based on the stated appropriate value in order to search for optimized process parameters.

2. The method for predicting wet process parameters in solar cell manufacturing according to claim 1, characterized in that, The process of obtaining the initial population based on the process parameters includes: The limitation ranges of the purification solution parameters, the etching solution parameters, and the cleaning solution parameters are obtained respectively; Obtain the number of individuals in the initial population; Based on the number of individuals, initial individuals of the purification solution parameters, the etching solution parameters, and the cleaning solution parameters are generated at equal intervals within the restricted range to obtain an initial population.

3. The method for predicting wet process parameters in solar cell manufacturing according to claim 2, characterized in that, The creation of a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-processing prediction model corresponding to the post-processing process includes: Construct a training dataset; wherein the training dataset includes process parameters and corresponding measurement data of silicon wafer surface cleanliness, cleaning fluid residue, silicon wafer purity, etching depth, etching residue, silicon wafer product cleanliness, and silicon wafer product flatness. The first preset neural network model is trained based on the training dataset to obtain the preprocessed prediction model; The second preset neural network model is trained based on the training dataset to obtain the purified prediction model; The etching prediction model is obtained by training the third preset neural network model based on the training dataset. The fourth preset neural network model is trained based on the training dataset to obtain the post-processing prediction model.

4. The method for predicting wet process parameters in solar cell manufacturing according to claim 3, characterized in that, The step of training the first preset neural network model based on the training dataset to obtain the preprocessed prediction model includes: The cleanliness measurement data and the cleaning fluid residue prediction parameters in the training dataset are used as the first training labels, and the silicon wafer type and the cleaning fluid parameters in the process parameters are used as the first training samples. The first training sample is input into the first preset neural network model to obtain the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters; Based on the first training label, the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters, the first preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the preprocessed prediction model. The step of training the second preset neural network model based on the training dataset to obtain the refined prediction model includes: The silicon wafer purity measured in the training dataset is used as the second training label, and the cleaning fluid parameter prediction data, the cleaning fluid parameter prediction parameters, and the purification fluid parameters in the process parameters are used as the second training samples. The second training sample is input into the second preset neural network model to obtain the predicted purity of the silicon wafer; Based on the second training label and the predicted purity of the silicon wafer, the second preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the purification prediction model. The step of training the third preset neural network model based on the training dataset to obtain the etching prediction model includes: The cleanliness measurement data in the training dataset is used as the third training label, and the predicted purity of the silicon wafer and the etching solution parameters in the process parameters are used as the third training samples. The third training sample is input into the third preset neural network model to obtain the etching prediction depth and etching residue prediction data; Based on the third training label, the etching prediction depth and the etching residue prediction data, the third preset neural network model is iteratively updated until the iteration termination condition is reached to obtain the etching prediction model. The step of training the fourth preset neural network model based on the training dataset to obtain the post-processing prediction model includes: The silicon wafer product cleanliness and silicon wafer product flatness in the training dataset are used as the fourth training labels, and the etching prediction depth, etching residue prediction data and cleaning fluid parameters in the process parameters are used as the fourth training samples. The fourth training sample is input into the fourth preset neural network model to obtain the predicted cleanliness and predicted flatness of the silicon wafer product. Based on the fourth training label, the predicted cleanliness of the silicon wafer product, and the predicted flatness of the silicon wafer product, the fourth preset neural network model is iteratively updated until the iteration termination condition is reached, thus obtaining the post-processing prediction model.

5. A storage medium, characterized in that, The storage medium stores a computer program, which is loaded by a processor to execute the steps in the wet process parameter prediction method for solar cell manufacturing according to any one of claims 1 to 4.

Images