A method, device, equipment and medium for standardizing execution of a single crystal pulling process
By constructing a golden process curve and a dynamic process window, the problem of the disconnect between process standards and production caused by static SOP in the preparation of monocrystalline silicon materials was solved. This enabled refined control and quality stability of the monocrystalline pulling process, reduced reliance on manual experience, and improved production efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SICHUAN GOKIN SOLAR TECHNOLOGY CO LTD
- Filing Date
- 2026-04-01
- Publication Date
- 2026-06-05
AI Technical Summary
In the current process of preparing monocrystalline silicon materials, the management of monocrystalline pulling process relies on static standard operating procedures (SOPs) and cannot be dynamically adjusted. This leads to a disconnect between process standards and actual production, a lack of quantifiable process curves, reliance on manual experience, a lack of closed-loop process improvement, difficulty in identifying potential quality risks, and low production efficiency.
By using big data analytics to construct quantified gold process curves and dynamic process windows, the trajectory determination of process parameters throughout the crystal pulling process is realized, and a multi-level intelligent alarm and SOP adaptive iteration mechanism is established to improve the level of refined management and control.
It enables refined control of the single crystal pulling process, reduces reliance on manual experience, identifies potential quality risks, ensures product quality stability and consistency, dynamically matches changes in the production environment, and improves production efficiency.
Smart Images
Figure CN122155534A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of photovoltaic monocrystalline silicon material preparation and production quality control technology. Specifically, it relates to a standardized execution method, apparatus, equipment and medium for monocrystalline silicon pulling process. Background Technology
[0002] In the preparation of photovoltaic monocrystalline silicon materials, the process management of monocrystalline pulling is a core link in ensuring product quality and improving production efficiency. Currently, the industry's monocrystalline pulling process management mainly relies on static paper or electronic standard operating procedures (SOPs). The centralized control system in the production process only monitors safety by setting fixed upper and lower limits for parameters. This technical model has exposed many defects in actual production and can no longer meet the requirements of high-precision and high-efficiency monocrystalline silicon production.
[0003] Static SOPs with poor adaptability: Existing SOPs are fixed documents that cannot be dynamically adjusted according to actual working conditions such as changes in equipment operating status, thermal aging, and raw material quality fluctuations during the production process, resulting in a disconnect between process standards and actual production. The single judgment dimension makes it difficult to identify potential risks: The centralized control system only adopts a simple parameter threshold alarm mechanism, which can only judge whether a single parameter exceeds a fixed upper or lower limit. It cannot make in-depth judgments on the continuous trajectory consistency of core process parameters such as pulling speed, heating power, and furnace temperature. It is difficult to discover potential quality risks that parameters are within the threshold range but deviate from the optimal path, which can easily lead to subsequent product quality problems. Over-reliance on human experience makes it difficult to pass on the technology: The production process lacks a quantifiable golden process curve as an operational benchmark. Process optimization and operation standards mainly rely on the human experience of core technical personnel. New employees cannot intuitively understand the gap between the current operation and the optimal operation through the system, and human experience is difficult to accurately replicate and pass on. The lack of a closed-loop process improvement mechanism and the lag in SOP iteration: The massive amount of process data generated during production is not effectively utilized, and production data is disconnected from process improvement. The upgrading and optimization of SOPs rely solely on manual post-event review, lacking automated self-learning and self-iteration mechanisms, resulting in process standard updates lagging behind actual production needs. Summary of the Invention
[0004] In view of this, the purpose of this application is to provide a standardized execution method, device, equipment and medium for the single crystal pulling process, which can construct a quantitative golden process curve through big data analysis, realize the trajectory determination of process parameters throughout the crystal pulling process, multi-level intelligent alarm, and establish an SOP adaptive iteration mechanism to improve the level of refined control of the single crystal pulling process and reduce the dependence on human experience.
[0005] In a first aspect, this application provides a standardized execution method for a single crystal pulling process, the method comprising the following steps: Collect and preprocess the crystal pulling data from different production bases and for different products during the crystal pulling process, and establish a crystal pulling dataset. Based on the crystal pulling dataset, a gold process curve and a dynamic process window are generated; The real-time release data is compared with the generated gold process curve and dynamic process window to generate a trajectory consistency judgment result, and multi-level alarms are triggered based on the judgment result. Regularly check the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve deviates systematically and the deviation exceeds the set ratio, trigger the standard operating procedure (SOP) update.
[0006] In some embodiments, distributed data collection is carried out on the complete production cycle of each crystal pulling furnace for monocrystalline silicon products of different production bases and different specifications. The collection object is the full cycle data of monocrystalline pulling from the welding step, through crystal pulling, shoulder setting, shoulder turning, equal diameter to the end and the completion of rod taking.
[0007] In some embodiments, generating the gold process curve and dynamic process window based on the crystal pulling dataset includes the following steps: A multi-indicator weighted scoring model is constructed. Scoring indicators covering two dimensions, production efficiency and product quality, are selected, and the weights of each scoring indicator are determined by a combination of the analytic hierarchy process and the entropy weighting method. The comprehensive score results of each furnace are obtained based on the constructed multi-index weighted scoring model. Based on the comprehensive score results, high-quality furnaces are selected and their corresponding ignition data are extracted as the original dataset for constructing the gold process curve. After feature alignment of all time-series data in the original dataset using the dynamic time warping algorithm, the mean curve of each parameter is calculated, and this mean curve is defined as the gold process curve. The standard deviation of each parameter is also calculated, and a dynamic process window is generated with the mean as the center and the standard deviation as the floating range.
[0008] In some embodiments, the scoring formula of the multi-index weighted scoring model is as follows:
[0009] in, Indicates the yield rate. Indicates the yield rate. Indicates the single furnace capacity. Indicates product quality; , , , These represent the weights of the corresponding scoring indicators.
[0010] In some embodiments, comparing the real-time release data with the generated gold process curve and dynamic process window includes the following steps: Real-time release data during the single crystal pulling process is collected synchronously, and the cumulative deviation area of the real-time release data curve relative to the gold process curve is calculated. By combining the boundary constraints of the dynamic process window, trajectory consistency is determined by the cumulative deviation area.
[0011] In some embodiments, the periodic detection of the deviation between the actual optimal process curve and the gold process curve, and if the center value of the actual optimal process curve undergoes a systematic shift and the shift exceeds a set proportion, triggering a standard operating procedure (SOP) update, includes the following steps: Acquire all the release data within the current set time period according to the preset cycle, and obtain the actual optimal process curve based on the release data; The actual optimal process curve is compared with the currently used gold process curve parameter by parameter, the center value offset of each parameter curve is calculated, and compared with the preset ratio threshold to determine whether there is a systematic offset. If a systematic offset occurs, regenerate the gold process curve and dynamic process window, and update the SOP based on the regenerated gold process curve and dynamic process window.
[0012] In some embodiments, the method further includes the following steps: The gold processing curve, dynamic process window, and real-time release data are transformed into a visual navigation map for display.
[0013] Secondly, this application also provides a standardized execution device for a single crystal pulling process, the device comprising: The data acquisition module is used to collect the release data during the single crystal pulling process of different production bases and different products, and to preprocess the data to establish a crystal pulling dataset. The module is used to generate gold process curves and dynamic process windows based on the crystal pulling dataset; The judgment module is used to compare the real-time release data with the generated gold process curve and dynamic process window, generate a trajectory consistency judgment result, and trigger multi-level alarms based on the judgment result. The update module is used to periodically detect the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve deviates systematically and the deviation exceeds the set proportion, the standard operating procedure (SOP) update is triggered.
[0014] Thirdly, this application also provides an electronic device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the steps of the standardized execution method for a single crystal pulling process described in any of the first aspects are executed.
[0015] Fourthly, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the standardized execution method for a single crystal pulling process described in any one of the first aspects.
[0016] This application describes a standardized execution method, apparatus, equipment, and medium for monocrystalline silicon pulling processes. It collects lead-in data from different production bases and for different products during the monocrystalline silicon pulling process, preprocesses it, and establishes a pulling dataset. Based on the pulling dataset, it generates a gold process curve and a dynamic process window. It compares real-time lead-in data with the generated gold process curve and dynamic process window to generate a trajectory consistency judgment result, and triggers multi-level alarms based on this result. It periodically checks the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve systematically shifts and the shift exceeds a set proportion, it triggers a standard operating procedure (SOP) update. This achieves dynamic matching between process standards and actual production, ensuring the stability and consistency of monocrystalline silicon product quality. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 A flowchart of the standardized execution method for the single crystal pulling process described in the embodiments of this application is shown; Figure 2 This document illustrates a flowchart of an embodiment of the present application that generates a gold process curve and a dynamic process window based on the crystal pulling dataset. Figure 3 A flowchart illustrating the triggering of standard operating procedure (SOP) updates according to an embodiment of this application is shown; Figure 4 A schematic diagram of the structure of the standardized execution device for the single crystal pulling process described in an embodiment of this application is shown; Figure 5 A schematic diagram of the structure of the electronic device described in an embodiment of this application is shown. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.
[0020] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.
[0021] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.
[0022] To address the shortcomings of existing technologies, this application provides a standardized execution method, apparatus, equipment, and medium for the single crystal pulling process. It can construct a quantified golden process curve through big data analysis, realize the trajectory determination of process parameters throughout the crystal pulling process, implement multi-level intelligent alarms, and establish an SOP adaptive iteration mechanism to improve the level of refined control of the single crystal pulling process and reduce reliance on human experience.
[0023] See the instruction manual appendix Figure 1 This application provides a standardized execution method for a single crystal pulling process, the method comprising the following steps: S1. Collect the lead-in and release data during the single crystal pulling process of different production bases and different products, and preprocess the data to establish a crystal pulling dataset; S2. Generate a gold process curve and a dynamic process window based on the crystal pulling dataset; S3. Compare the real-time release data with the generated gold process curve and dynamic process window to generate a trajectory consistency judgment result, and trigger multi-level alarms based on the judgment result. S4. Periodically check the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve deviates systematically and the deviation exceeds the set ratio, trigger the standard operating procedure (SOP) update.
[0024] Specifically, step S1 mainly involves obtaining standardized, high-quality basic data for the entire single crystal pulling process, providing a data source and foundation for subsequent process analysis, benchmark construction, and real-time judgment, thus solving the problem of the lack of unified quantitative data support for process analysis in existing technologies.
[0025] Specifically, for monocrystalline silicon products of different production bases and specifications, distributed data collection is carried out on the complete production cycle of each crystal pulling furnace. The data collection targets the entire cycle of monocrystalline pulling, from the welding step, through crystal pulling, shoulder setting, shoulder turning, equal diameter setting to the final stage and rod removal. During the collection process, each crystal pulling furnace is equipped with an independent data collection node to ensure the independence and accuracy of data collection. The collection frequency matches the production process rhythm to ensure the continuity of time-series data.
[0026] Furthermore, the collected raw release data is preprocessed to remove outliers caused by equipment malfunctions or abnormal acquisition terminals, missing values due to acquisition interruptions, and duplicate values resulting from repeated acquisitions. Valid and continuous raw data is retained to avoid interference from invalid data in subsequent analysis. Additionally, the cleaned valid data is standardized, including data units, data formats, and time axis calibration, ensuring the comparability and analyzability of release data from different production bases, different crystal pulling furnaces, and different products. All standardized data is then integrated and stored to establish a structured and callable crystal pulling dataset, providing a fundamental data source for constructing gold process curves.
[0027] Step S2 mainly involves extracting the optimal process quantification benchmark (gold process curve and dynamic process window) from massive historical crystal pulling data, transforming vague manual process experience into quantifiable and comparable objective process trajectory standards, and solving the problem of the lack of a unified optimal process benchmark and reliance on manual experience in existing technologies.
[0028] See the instruction manual appendix Figure 2 The process of generating gold process curves and dynamic process windows based on the crystal pulling dataset includes the following steps: S201. Construct a multi-index weighted scoring model; wherein, scoring indicators covering two dimensions, production efficiency and product quality, are selected, and the weight of each scoring indicator is determined by a combination of the analytic hierarchy process and the entropy weight method. S202. Based on the constructed multi-index weighted scoring model, the comprehensive scoring results of each furnace are obtained, and high-quality furnaces are selected according to the comprehensive scoring results, and their corresponding ignition and release data are extracted as the original dataset for constructing the gold process curve. S203. After aligning the features of all time-series data in the original dataset using the dynamic time warping algorithm, calculate the mean curve of each parameter, define the mean curve as the gold process curve, and calculate the standard deviation of each parameter. With the mean as the center and the standard deviation as the floating range, generate a dynamic process window.
[0029] Before executing step S201, all valid historical data are retrieved from the structured crystal pulling dataset established in step S1 as the original data material for constructing the gold process curve. During the retrieval process, the furnace identification, product type, production base and other related information of the data are retained to facilitate subsequent data screening and traceability.
[0030] In steps S201 and S202, when constructing the multi-index weighted scoring model, crystal formation rate, yield, single-furnace capacity, and product quality are selected as core scoring indicators, covering the two core dimensions of production efficiency and product quality. A combined weighting method using the analytic hierarchy process (AHP) and entropy weighting is employed to determine the weights of each indicator. Subjective weights for crystal formation rate and yield are set based on production targets, while objective weights for single-furnace capacity and product quality are automatically allocated based on the dispersion of historical data. Then, the single-crystal pulling output of each furnace is quantitatively scored according to the scoring formula to obtain the comprehensive score result for each furnace. The scoring formula for the multi-index weighted scoring model is as follows:
[0031] in, Indicates the yield rate. Indicates the yield rate. Indicates the single furnace capacity. Indicates product quality; , , , These represent the weights of the corresponding scoring indicators. Then, the comprehensive scores of all furnace runs are sorted from highest to lowest, and the top 10% of furnace runs are selected as high-quality furnace runs. Their corresponding induction and release data are extracted as the original dataset for constructing the gold process curve.
[0032] In step S203, firstly, to address the issue of varying process durations among different furnaces in the high-quality furnace dataset due to subtle differences in production conditions, a dynamic time warping algorithm is used to align the features of all time-series data, eliminating differences in the time dimension and ensuring that the process parameters of each furnace are comparable and calculable on a unified time axis. Then, on the aligned unified time axis, each core process parameter is calculated separately, and the mean curve of each parameter is defined as the golden process curve, serving as the optimal process trajectory benchmark for single crystal pulling. Furthermore, the standard deviation of each parameter is calculated, and a dynamic process window is generated with the mean as the center and the standard deviation as the floating range, serving as a reasonable floating boundary for the process parameters. For the process characteristics of each step in single crystal pulling, a dedicated golden process curve and dynamic process window are generated, and all benchmark data are stored in the process benchmark library.
[0033] Step S3 mainly involves comparing the real-time production data of single crystal pulling with the optimal process benchmark in a trajectory manner, thereby upgrading from single-point threshold judgment to full-process trajectory control, identifying potential quality risks in advance and issuing alarms according to risk level, and solving the problems of single judgment dimension, difficulty in discovering potential risks, and delayed alarms in existing technologies.
[0034] In one embodiment, firstly, real-time release data during the single crystal pulling process is collected synchronously, with the data dimension and collection frequency consistent with the release data in step S1. The collected real-time data is used as the real-time object for trajectory consistency determination. Then, the gold process curve benchmark value and dynamic process window upper and lower limits for the corresponding crystal pulling step are retrieved from the process benchmark library. The real-time release data is mapped to the actual process curve in the time-parameter coordinate system, forming a double-line comparison with the gold process curve in the same coordinate system. Using 5 minutes as the minimum judgment time window, the trapezoidal area method is used to calculate the single-cycle deviation area of the actual process curve relative to the gold process curve within a single time window. The cumulative deviation area is obtained by superimposing consecutive cycles. Combined with the boundary constraints of the dynamic process window, the changing trend of the cumulative deviation area is analyzed. If the actual curve is within the dynamic process window, and the cumulative deviation area of 3 or more consecutive time windows shows a continuous increasing trend with a growth rate ≥ a preset threshold, it is determined that a trajectory drift phenomenon has occurred, realizing full trajectory quantification judgment of the crystal pulling process.
[0035] The system classifies real-time judgment results into risk levels based on the cumulative deviation area (value / proportion), whether trajectory drift is triggered, and whether the dynamic process window is exceeded. The classification criteria are precisely matched with multi-level alarm rules. For example, based on the risk level classification results, corresponding L1, L2, and L3 level alarms are triggered and preset processing actions are executed. Each level of alarm is triggered and handled independently, and alarm escalation for multi-parameter collaborative anomalies is supported. When the alarm is cleared, the system automatically resets and records relevant information.
[0036] L1 level (blue) slight deviation: The deviation area within a single time window is less than the preset basic threshold, or the cumulative deviation area does not reach the trajectory drift judgment standard, and the actual value is always within the dynamic process window. A blue slight deviation mark is displayed on the production dashboard, and the deviation data is automatically recorded to the database. Level L2 (Yellow) Trend Deviation: This is determined to be trajectory drift, or the actual value is within the dynamic process window but the single parameter deviation rate is >5%, or multiple process parameters show a slight deviation at Level L1 at the same time. Yellow warning information is pushed to process engineers and team leaders via DingTalk and WebHook. A yellow flashing icon is displayed on the production dashboard, requiring process engineers to check and judge whether the operating parameters need to be adjusted within a preset time. Level L3 (Red) Severe Deviation: The actual process parameters exceed the upper and lower limits of the dynamic process window, or the trajectory drift is not handled in time, resulting in a breach of the dynamic process window, or the single parameter deviation rate is ≥20%. An on-site audible and visual alarm will be triggered immediately, and a red emergency alarm window will pop up in the central control center. The system will automatically associate the corresponding crystal rod number and mark it as an "abnormal record", push emergency alarm information to the management personnel, and require immediate on-site intervention. If necessary, it will trigger protective actions such as automatic power reduction and shutdown.
[0037] Step S4 mainly establishes a dynamic self-evolution mechanism for process standards, enabling the gold process curve and SOP to be automatically optimized and updated according to changes in actual working conditions such as production environment, thermal aging, and raw material fluctuations, thus solving the problems of static SOPs, disconnection from actual production, and lack of closed-loop iteration mechanism in existing technologies.
[0038] See the instruction manual appendix Figure 3 The periodic detection of the deviation between the actual optimal process curve and the gold process curve, if the center value of the actual optimal process curve undergoes a systematic shift and the shift exceeds a set proportion, triggers a standard operating procedure (SOP) update, including the following steps: S401. Obtain all the release data within the current set time period according to the preset cycle, and obtain the actual optimal process curve based on the release data; S402. Compare the actual optimal process curve with the currently used gold process curve parameter by parameter, calculate the center value offset of each parameter curve, and compare it with the preset ratio threshold to determine whether there is a systematic offset. S403. If a systematic offset occurs, regenerate the gold process curve and dynamic process window, and update the SOP based on the regenerated gold process curve and dynamic process window.
[0039] Specifically, firstly, following a monthly iteration cycle, the induction and release data of all furnaces within the past 30 days are extracted from the crystal pulling dataset updated in real time in step S1. The multi-index weighted scoring and high-quality furnace selection process in step S2 is repeated to obtain the actual process curve of the recent high-quality furnaces, which serves as the actual optimal process curve. Then, the original gold process curve and dynamic process window are retrieved from the process benchmark library. The actual optimal process curve is compared with the original gold process curve parameter by parameter in a refined manner, and the center value offset of each core process parameter curve is calculated, recording the offset direction and magnitude. Among them, the center value offset of each parameter is compared with a preset proportion threshold. If it is determined that the center value of the actual optimal process curve has undergone a systematic shift and the offset exceeds the set proportion, it is considered that the operating conditions such as the production environment, thermal field, and raw materials have changed, triggering the SOP update command; if the offset threshold is not reached, the original gold process curve, dynamic process window, and SOP remain unchanged.
[0040] Specifically, based on the SOP update instruction, the time-series data alignment, mean, and standard deviation calculation process in step S2 is re-executed. Based on the most recent optimal process curve, a new golden process curve and a new dynamic process window are regenerated for each process step. Based on the new golden process curve and dynamic process window, the entire system of digital SOP content, including operation specifications, judgment criteria, and alarm thresholds, is updated to generate a new version of SOP. The new version of SOP is then synchronized to the process benchmark library, replacing the original version, achieving a seamless upgrade of process standards and ensuring that subsequent trajectory consistency judgments, multi-level alarms, and visualization displays all use the latest benchmark. At the same time, the system automatically retains complete traceability records, including the reason for the SOP update, update time, offset parameters and magnitude, and affected production bases / product types, achieving full traceability of process standard iteration.
[0041] Furthermore, this application provides a standardized execution method for a single crystal pulling process, the method further comprising the following steps: S5. Transform the gold process curve, dynamic process window, and real-time pilot-release data into a visual navigation map for display.
[0042] Step S5 primarily transforms abstract process curves and data deviations into an intuitive visual interface, lowering the operational threshold for operators and enabling them to quickly perceive the deviation between the current operation and the optimal process. This ensures that the process benchmarks and real-time judgment results from previous steps can be applied to actual production operations, solving the problem that existing technologies use text-based SOPs and make it difficult to intuitively perceive operational deviations.
[0043] Specifically, the process first retrieves the gold process curve, dynamic process window, and benchmark data corresponding to the latest version of SOP generated in step S2 from the process benchmark library, as well as real-time crystal pulling data, real-time trajectory deviation results, and alarm information. All multi-source data are integrated to ensure the comprehensiveness, real-time nature, and accuracy of the displayed data. Then, the integrated abstract data is transformed into a digital visualization navigation map, using a differentiated color scheme to display the gold process curve and the actual real-time process curve in curve form, and the dynamic process window in shaded area form. A striking color rule is set to highlight the real-time curve portion that deviates from the optimal process path, and visual pop-up prompts are provided for alarm information. The completed visualization navigation map is then simultaneously displayed on the crystal pulling furnace's on-site operation terminal and the central control center monitoring terminal, achieving multi-terminal synchronous display of process benchmarks and real-time production data. Operators can intuitively perceive the deviation position and magnitude between the current operation and the optimal process path through the visual interface, view historical deviation records and alarm information, and adjust operating parameters promptly based on the visual deviation prompts. This achieves efficient interaction between process standards and actual production operations, ensuring standardized process execution.
[0044] This application provides a standardized execution method for the single-crystal pulling process. On one hand, by constructing a golden process curve and a dynamic process window, it transforms vague manual operation experience into a quantifiable process trajectory benchmark. Combined with the cumulative deviation area, it achieves a quantitative determination of the consistency of continuous process parameter trajectories, breaking through the limitations of traditional single threshold determination. It can accurately identify trajectory drift phenomena within the dynamic process window, providing an objective and accurate quantitative basis for process determination and eliminating the subjectivity and randomness of manual experience-based judgment. On the other hand, it establishes a three-level alarm mechanism based on trajectory consistency determination results. Differentiated triggering conditions and handling actions are set for different degrees of process deviation. It can identify potential risks and issue early warnings at the hidden stage before quality defects occur, realizing the transformation from post-remediation to pre-prevention. At the same time, through multi-parameter collaborative abnormal alarm escalation and protective action triggering, it effectively avoids serious quality risks and reduces product defect rates. Thirdly, an adaptive iteration mechanism for SOPs based on production big data was established. This mechanism automatically detects the deviation of the actual optimal process curve on a monthly basis. When the production conditions undergo systematic changes, the golden process curve and dynamic process window are automatically regenerated, the content of the entire SOP system is updated, and traceability records are retained. This solves the problems of static SOPs and disconnection from actual production. It achieves dynamic matching between process standards and changes in production environment, thermal aging, raw material fluctuations, and other operating conditions, forming a closed loop for process improvement.
[0045] As per the instruction manual Figure 4 As shown in the embodiment of this application, a standardized execution device for a single crystal pulling process is also provided, the device comprising: The acquisition module 401 is used to collect the lead-in data during the single crystal pulling process of different production bases and different products, and to preprocess the data to establish a crystal pulling dataset. Module 402 is used to generate gold process curves and dynamic process windows based on the crystal pulling dataset; The judgment module 403 is used to compare the real-time release data with the generated gold process curve and dynamic process window, generate a trajectory consistency judgment result, and trigger multi-level alarms based on the judgment result. The update module 404 is used to periodically detect the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve is systematically deviated and the deviation exceeds the set ratio, the standard operating procedure (SOP) update is triggered.
[0046] In some embodiments, the acquisition module 401 performs distributed data acquisition on the complete production cycle of each crystal pulling furnace for monocrystalline silicon products of different production bases and different specifications. The acquisition object is the full cycle data of monocrystalline pulling from the welding step, through crystal pulling, shoulder setting, shoulder turning, equal diameter to the end and the completion of rod taking.
[0047] In some embodiments, the construction module 402 generates a gold process curve and a dynamic process window based on the crystal pulling dataset, including: constructing a multi-index weighted scoring model; wherein, scoring indicators covering two dimensions, production efficiency and product quality, are selected, and a combined weighting method combining the analytic hierarchy process (AHP) and entropy weighting is used to determine the weight of each scoring indicator; based on the constructed multi-index weighted scoring model, a comprehensive scoring result for each furnace is obtained, and high-quality furnaces are selected according to the obtained comprehensive scoring results, and their corresponding induction and release data are extracted as the original dataset for constructing the gold process curve; after feature alignment of all time-series data in the original dataset using a dynamic time warping algorithm, the mean curve of each parameter is calculated, the mean curve is defined as the gold process curve, and the standard deviation of each parameter is calculated, with the mean as the center and the standard deviation as the floating range, to generate a dynamic process window.
[0048] In some embodiments, the determination module 403 compares the real-time release data with the generated gold process curve and the dynamic process window, including: synchronously collecting real-time release data during the single crystal pulling process, calculating the cumulative deviation area of the real-time release data curve relative to the gold process curve; and combining the boundary constraints of the dynamic process window to determine the trajectory consistency through the cumulative deviation area.
[0049] In some embodiments, the update module 404 periodically detects the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve undergoes a systematic deviation and the deviation exceeds a set proportion, the standard operating procedure (SOP) is updated, including: acquiring all release data within the current set time period according to a preset cycle, and obtaining the actual optimal process curve based on the release data; comparing the actual optimal process curve with the currently used gold process curve parameter by parameter, calculating the center value deviation of each parameter curve, and comparing it with a preset proportion threshold to determine whether there is a systematic deviation; if a systematic deviation occurs, regenerating the gold process curve and the dynamic process window, and updating the SOP based on the regenerated gold process curve and the dynamic process window.
[0050] In some embodiments, the apparatus further includes: The display module is used to transform the gold process curve, dynamic process window, and real-time release data into a visual navigation map for display.
[0051] The standardized execution device for the monocrystalline silicon pulling process described in this application collects lead-in data from different production bases and for different products during the monocrystalline silicon pulling process using a data acquisition module. This data is then preprocessed to create a pulling dataset. A construction module generates a gold process curve and a dynamic process window based on this dataset. A judgment module compares the real-time lead-in data with the generated gold process curve and dynamic process window to generate a trajectory consistency judgment result, triggering multi-level alarms based on this result. An update module periodically checks the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve experiences a systematic shift exceeding a set proportion, a standard operating procedure (SOP) update is triggered. This achieves dynamic matching between process standards and actual production, ensuring the stability and consistency of monocrystalline silicon product quality.
[0052] Based on the same concept of the present invention, as shown in the appendix to the specification. Figure 5 As shown in the embodiment of this application, an electronic device 500 is provided. The electronic device 500 includes: at least one processor 501, at least one network interface 504 or other user interface 503, a memory 505, and at least one communication bus 502. The communication bus 502 is used to enable communication between these components. The electronic device 500 may optionally include a user interface 503, including a display (e.g., touchscreen, LCD, CRT, holographic imaging, or projector), a keyboard, or a clicking device (e.g., mouse, trackball, touchpad, or touchscreen).
[0053] Memory 505 may include read-only memory and random access memory, and provides instructions and data to processor 501. A portion of memory 505 may also include non-volatile random access memory (NVRAM).
[0054] In some implementations, memory 505 stores executable modules or data structures, or subsets thereof, or extended sets thereof: The 5051 operating system contains various system programs used to implement various basic business functions and handle hardware-based tasks. Application module 5052 contains various applications, such as launchers, media players, and browsers, to implement various application functions.
[0055] In this embodiment of the application, the processor 501 executes steps such as a standardized execution method for a single crystal pulling process by calling a program or instruction stored in the memory 505.
[0056] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor, such as steps in a standardized execution method for a single crystal pulling process.
[0057] Specifically, the storage medium can be a general-purpose storage medium, such as a portable disk or hard drive. When the computer program on the storage medium is run, it achieves dynamic matching between process standards and actual production, ensuring the stability and consistency of the quality of monocrystalline silicon products.
[0058] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, and the indirect coupling or communication connection of the apparatus or units may be electrical, mechanical, or other forms.
[0059] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0060] In addition, the functional units in the embodiments provided in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0061] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0062] Finally, it should be noted that the above embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.
Claims
1. A standardized execution method for a single crystal pulling process, characterized in that, The method includes the following steps: Collect and preprocess the crystal pulling data from different production bases and for different products during the crystal pulling process, and establish a crystal pulling dataset. Based on the crystal pulling dataset, a gold process curve and a dynamic process window are generated; The real-time release data is compared with the generated gold process curve and dynamic process window to generate a trajectory consistency judgment result, and multi-level alarms are triggered based on the judgment result. Regularly check the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve deviates systematically and the deviation exceeds the set ratio, trigger the standard operating procedure (SOP) update.
2. The standardized execution method for a single crystal pulling process according to claim 1, characterized in that, in, For monocrystalline silicon products of different production bases and specifications, distributed data collection is carried out on the complete production cycle of each crystal pulling furnace. The data collection object is the full cycle of monocrystalline pulling from the welding step, through crystal pulling, shoulder setting, shoulder turning, equal diameter setting to the end and the completion of the rod taking.
3. The standardized execution method for a single crystal pulling process according to claim 2, characterized in that, The process of generating the gold process curve and dynamic process window based on the crystal pulling dataset includes the following steps: A multi-indicator weighted scoring model is constructed. Scoring indicators covering two dimensions, production efficiency and product quality, are selected, and the weights of each scoring indicator are determined by a combination of the analytic hierarchy process and the entropy weighting method. The comprehensive score results of each furnace are obtained based on the constructed multi-index weighted scoring model. Based on the comprehensive score results, high-quality furnaces are selected and their corresponding ignition data are extracted as the original dataset for constructing the gold process curve. After feature alignment of all time-series data in the original dataset using the dynamic time warping algorithm, the mean curve of each parameter is calculated, and this mean curve is defined as the gold process curve. The standard deviation of each parameter is also calculated, and a dynamic process window is generated with the mean as the center and the standard deviation as the floating range.
4. The standardized execution method for a single crystal pulling process according to claim 3, characterized in that, The scoring formula for the multi-index weighted scoring model is as follows: in, Indicates the yield rate. Indicates the yield rate. Indicates the single furnace capacity. Indicates product quality; , , , These represent the weights of the corresponding scoring indicators.
5. The standardized execution method for a single crystal pulling process according to claim 4, characterized in that, The comparison of real-time release data with the generated gold process curve and dynamic process window includes the following steps: Real-time release data during the single crystal pulling process is collected synchronously, and the cumulative deviation area of the real-time release data curve relative to the gold process curve is calculated. By combining the boundary constraints of the dynamic process window, trajectory consistency is determined by the cumulative deviation area.
6. The standardized execution method for a single crystal pulling process according to claim 5, characterized in that, The periodic detection of the deviation between the actual optimal process curve and the gold process curve, if the center value of the actual optimal process curve deviates systematically and the deviation exceeds a set proportion, triggers a standard operating procedure (SOP) update, including the following steps: Acquire all the release data within the current set time period according to the preset cycle, and obtain the actual optimal process curve based on the release data; The actual optimal process curve is compared with the currently used gold process curve parameter by parameter, the center value offset of each parameter curve is calculated, and compared with the preset ratio threshold to determine whether there is a systematic offset. If a systematic offset occurs, regenerate the gold process curve and dynamic process window, and update the SOP based on the regenerated gold process curve and dynamic process window.
7. The standardized execution method for a single crystal pulling process according to claim 6, characterized in that, The method further includes the following steps: The gold processing curve, dynamic process window, and real-time release data are transformed into a visual navigation map for display.
8. A standardized execution device for a single crystal pulling process, characterized in that, The device includes: The data acquisition module is used to collect the release data during the single crystal pulling process of different production bases and different products, and to preprocess the data to establish a crystal pulling dataset. The module is used to generate gold process curves and dynamic process windows based on the crystal pulling dataset; The judgment module is used to compare the real-time release data with the generated gold process curve and dynamic process window, generate a trajectory consistency judgment result, and trigger multi-level alarms based on the judgment result. The update module is used to periodically detect the deviation between the actual optimal process curve and the gold process curve. If the center value of the actual optimal process curve deviates systematically and the deviation exceeds the set proportion, the standard operating procedure (SOP) update is triggered.
9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of a standardized execution method for a single crystal pulling process as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of a standardized execution method for a single crystal pulling process as described in any one of claims 1 to 7.