System and method for processing surface scratch data of zinc-aluminum-magnesium steel strip

Abstract

The application discloses a kind of zinc-aluminum-magnesium steel strip surface scratch data processing system and method, it is related to metal surface scratch data processing technical field.The system includes material parameter input module, original scratch data input module, zinc-aluminum-magnesium special feature database, multi-source parameter correlation learning module, double dynamic verification module, invalid data filtering module, two-way feedback correction module and effective data output and digital twin module.The present application constructs the parameter verification and data processing system of zinc-aluminum-magnesium speciality, accurately filters pseudo scratch and invalid data, avoids the problem of poor adaptability of general scheme, reaches the effect of improving scratch data processing precision;Through the two-way feedback of data processing and front-end equipment, the quality of original data acquisition is optimized from the source, and the generation of subsequent invalid data is reduced, to ensure the effect of production data stability;Through digital twin body correlation whole-chain data and backwash processing rule, realize the closed-loop optimization of data from processing to application.

Description

Technical Field

[0001] This invention relates to the field of metal surface scratch data processing technology, specifically to a system and method for processing scratch data on zinc-aluminum-magnesium steel strip surfaces. Background Technology

[0002] Zinc-aluminum-magnesium steel strip, an alloy-coated steel containing zinc, magnesium, and aluminum, is widely used in automobile manufacturing, home appliance housings, and pipeline engineering due to its high corrosion resistance and strength. Its surface quality directly determines the corrosion resistance and appearance pass rate of the product. Scratches, a common defect in the production, transportation, and processing of this type of steel strip, make surface quality control a critical requirement in industrial production. Currently, the industry relies heavily on standardized data processing solutions for handling metal surface scratches. These solutions typically use fixed thresholds (such as uniform grayscale ranges and scratch length / depth standards) for data screening and judgment, without customizing the design to suit the material characteristics of zinc-aluminum-magnesium steel strip. Because of its magnesium content, zinc-aluminum-magnesium coatings are prone to forming oxide spots during storage or production. These spots have grayscale values ​​very close to those of shallow, effective scratches, and differences in parameters such as coating thickness (5-15μm) and magnesium content (1.5%-3.0%) directly affect the presentation of scratch characteristics. Standardized solutions cannot dynamically adapt to these changes in material characteristics.

[0003] The core flaw of existing technology lies in the disconnect between the generalized data processing solution and the specific material characteristics of zinc-aluminum-magnesium steel strip, leading to two key problems: First, oxide spots are easily misjudged as valid scratches, resulting in a high false scratch detection rate and increasing unnecessary quality review costs. Second, fine and shallow scratches that meet the quality requirements of zinc-aluminum-magnesium steel strip (such as depth <0.05mm and length <0.3mm) are often mistakenly classified as noise data and rejected because they do not match the specific judgment rules, resulting in the failure to detect valid scratches. This ultimately leads to batch quality judgment deviations, which may cause qualified products to be mistakenly judged as unqualified and wasted, or allow products with potential quality risks to enter the market, affecting the performance of downstream applications. Summary of the Invention

[0004] The purpose of this invention is to provide a data processing system and method for surface scratches on zinc-aluminum-magnesium steel strips to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a data processing system for surface scratches on zinc-aluminum-magnesium steel strip, comprising:

[0006] The material parameter input module is used to receive and verify the basic parameters of zinc-aluminum-magnesium steel strip and output the parameter package that has passed the verification.

[0007] The raw scratch data input module is used to receive the raw scratch data block output by the front-end detection device, perform integrity verification on the raw scratch data block, and output the raw data block that has passed the verification.

[0008] A dedicated feature database for zinc, aluminum, and magnesium is used to store dynamic feature parameters associated with the properties of zinc, aluminum, and magnesium steel strips. After receiving the parameter package, the database retrieves and outputs the corresponding feature judgment criteria.

[0009] The multi-source parameter association learning module is used to receive production process parameters and historical training data from the basic parameters and generate dynamic judgment thresholds;

[0010] The dual dynamic verification module receives the original data block that has passed verification, the feature judgment benchmark and the dynamic judgment threshold respectively, performs grayscale verification and morphological verification on the original data block that has passed verification, and outputs a data frame with verification result mark through key judgment logic.

[0011] The invalid data filtering module receives the data frame with the verification result mark, classifies the data frame with the verification result mark, and outputs valid data frames and device abnormal data.

[0012] The bidirectional feedback correction module receives the abnormal data from the equipment, analyzes the root cause of the abnormal data, and outputs equipment adjustment instructions.

[0013] The effective data output and digital twin module receives the effective data frame, outputs a quality report, and constructs a digital twin. The digital twin is used to associate the effective data frame, downstream processing data, and corrosion resistance data, and feeds back into the multi-source parameter association learning module.

[0014] Preferably, the material parameter input module includes:

[0015] The parameter receiving unit is used to receive the basic parameters of the zinc-aluminum-magnesium steel strip, including batch number, coating thickness, magnesium content, rolling completion time and production equipment number.

[0016] The parameter verification unit is used to verify the format, value range and validity of the basic parameters. If the verification fails, an exception code is output and a preset processing action is executed. The preset processing action includes refusing to receive or triggering manual confirmation.

[0017] The data interface unit uses a preset industrial Ethernet protocol to interface with the production execution system. The transmission rate of the preset industrial Ethernet protocol is not less than 100Mbps, and the parameter transmission delay is not more than 100ms.

[0018] The parameter package output unit is used to encapsulate the verified basic parameters into the parameter package and send it synchronously to the original scratch data input module and the zinc-aluminum-magnesium exclusive feature database.

[0019] Preferably, the parameter verification unit includes:

[0020] The format verification subunit is used to verify whether the format of the batch number is a preset identifier + 6-digit date + 3-digit serial number, wherein the preset identifier is "ZM" and the total length of the batch number is fixed at 12 digits;

[0021] The value verification subunit is used to verify whether the value range of the coating thickness is 5.0-15.0μm, whether the value range of the magnesium content is 1.5%~3.0%, whether the rolling completion time is not a future time, and whether the production equipment number is in the preset equipment list.

[0022] An exception handling subunit is used to output a corresponding exception code when the basic parameters do not meet the verification requirements of the format verification subunit or the value verification subunit. The exception codes include P001 for batch number format exception, P002 for coating thickness exception, P003 for magnesium content exception, P004 for rolling completion time exception, and P005 for production equipment number exception.

[0023] Preferably, the original scratch data input module includes:

[0024] The data block receiving unit is used to receive the original scratch data block sent by the front-end detection device. The original scratch data block is encapsulated in JSON format. Every 100 frames of data constitute one original scratch data block. The header of the original scratch data block contains a batch number, a data block number, and a verification field for the number of frames.

[0025] The integrity verification unit is used to verify whether the actual number of frames in the original scratch data block is consistent with the number of frames in the header, whether the detection area coordinates in the original scratch data block are not repeated, and whether the scratch depth and scratch length in the original scratch data block are non-negative values.

[0026] The data block processing unit is used to output the original data block that has passed the integrity verification unit to the dual dynamic verification module when the original scratch data block passes the verification; when the original scratch data block fails the verification, it triggers a retransmission instruction, the maximum number of times the retransmission instruction is triggered is 3. If the verification still fails after 3 retransmissions, the original scratch data block is marked as an abnormal data block and sent to the invalid data filtering module.

[0027] Preferably, the zinc-aluminum-magnesium specific feature database includes:

[0028] A three-dimensional index storage unit is used to store a tree-like index structure with coating thickness as the first-level index, magnesium content as the second-level index, and storage time as the third-level index. The storage time is divided into intervals of 0-24h, 24-48h, 48-72h, and greater than 72h.

[0029] The oxidation model unit has a built-in oxide layer thickness calculation model, and its calculation formula is as follows: ,in, This indicates the thickness of the oxide layer, in μm, with a value ranging from 0 to 2.0 μm. The rate constant is expressed in units of 1000 ppm. The correlation between the magnesium content and the magnesium content is positive, specifically: when the magnesium content is 1.5%, =0.02; when the magnesium content is 2.0%, =0.025; when the magnesium content is 2.5%, =0.03; when the magnesium content is 3.0%, =0.04; This represents the storage time, in hours, and ranges from 0 to 168 hours.

[0030] The feature benchmark output unit is used to receive the parameter package, calculate the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieve the three-dimensional index storage unit and call the oxidation model unit to correct the grayscale range, and output the feature judgment benchmark containing the grayscale range, grayscale compensation coefficient and depth-length rule to the dual dynamic verification module.

[0031] Preferably, the dual dynamic verification module includes:

[0032] A data integration unit is used to integrate the verified original data block, the feature determination benchmark, and the dynamic determination threshold.

[0033] The grayscale verification unit is used to calculate the storage time based on the collection timestamp in the verified original data block, call the oxidation model unit to calculate the oxide layer thickness to correct the grayscale range, and combine the light intensity to apply the grayscale compensation coefficient to determine whether the grayscale value in the verified original data block is within the corrected grayscale range.

[0034] The morphological verification unit is used to extract the scratch depth and scratch length from the original data block that has passed the verification, calculate the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and determine whether the ratio is not less than 1.

[0035] The key judgment unit is used to execute the key judgment logic to determine whether the original data block that has passed the verification has passed the verification of the grayscale verification unit and the morphological verification unit at the same time. If it passes at the same time, it is marked as data to be confirmed as valid data. If it does not pass at the same time, it is marked as invalid data to be classified. The unit outputs the data frame with the verification result mark containing the data to be confirmed as valid data and the data to be classified as invalid data.

[0036] Preferably, the effective data output and digital twin module includes:

[0037] The quality report generation unit is used to compile the valid data frames into a batch summary table and a list of valid scratch coordinates. The batch summary table includes the batch number, the total number of valid scratches, the number of scratches in different depth ranges, the maximum scratch size, and the data pass rate. The list of valid scratch coordinates includes the detection area coordinates, scratch depth, scratch length, and collection timestamp.

[0038] A digital twin construction unit is used to associate the batch summary table, downstream processing data, and corrosion resistance data to construct the digital twin. The downstream processing data includes scratch extension records in the stamping process, and the corrosion resistance data includes salt spray test results.

[0039] The data feedback unit is used to send the data in the digital twin to the multi-source parameter association learning module to trigger incremental training of the multi-source parameter association learning module and update the dynamic judgment threshold.

[0040] This invention also provides a method for processing surface scratch data of zinc-aluminum-magnesium steel strip, comprising the following steps:

[0041] S1: Material parameters and raw data are input synchronously; receive and verify the basic parameters of zinc-aluminum-magnesium steel strip; output the parameter package that has passed verification; receive the raw scratch data block output by the front-end detection equipment; after verifying the integrity of the raw scratch data block, output the raw data block that has passed verification.

[0042] S2: Dedicated feature parameter retrieval, receiving the parameter package, calculating the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieving preset dynamic feature parameters and correcting the grayscale range, and outputting the corresponding feature judgment benchmark;

[0043] S3: Multi-source parameter association rule generation: Receive production process parameters and historical training data from the basic parameters, train the production process parameters and historical training data using a preset algorithm, and generate a dynamic judgment threshold.

[0044] S4: Dual dynamic verification and key judgment, integrating the verified original data block, the feature judgment benchmark, and the dynamic judgment threshold, sequentially performing grayscale verification and morphological verification on the verified original data block, and outputting a data frame marked with the verification result through key judgment logic; wherein, the grayscale verification includes calculating the storage time based on the acquisition timestamp in the verified original data block, correcting the grayscale range by calling the oxide layer thickness calculation formula, applying a grayscale compensation coefficient in combination with the illumination intensity, and judging whether the grayscale value in the verified original data block is within the corrected grayscale range; the morphological verification includes extracting the scratch depth and scratch length in the verified original data block, calculating the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and judging whether the ratio is not less than 1;

[0045] S5: Invalid data filtering and classification: Receive the data frame with verification result mark, classify the data frame with verification result mark according to the preset classification rule, and output valid data frame and device abnormal data;

[0046] S6: Two-way feedback correction: Receive the abnormal data of the device, analyze the abnormal fields and associated detection device information corresponding to the abnormal data of the device, locate the root cause of the abnormality and output the device adjustment command;

[0047] S7: Valid data output and full lifecycle traceability. Receive the valid data frame, compile the valid data frame into a quality report, and construct a digital twin by associating the quality report, downstream processing data, and corrosion resistance data. The digital twin is used to feed back and update the dynamic judgment threshold.

[0048] The present invention also provides an electronic device, which is a physical device, comprising:

[0049] The processor and the memory are communicatively connected.

[0050] The memory is used to store at least one executable instruction executed by the processor, which executes the executable instruction to implement the zinc-aluminum-magnesium steel strip surface scratch data processing method as described above.

[0051] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for processing surface scratch data of zinc-aluminum-magnesium steel strip.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] By constructing a dedicated parameter verification and data processing system for zinc, aluminum, and magnesium, false scratches and invalid data are accurately filtered out, avoiding the problem of poor adaptability of general solutions, thus improving the accuracy of scratch data processing. Through two-way feedback between data processing and front-end equipment, the quality of raw data collection is optimized from the source, reducing the generation of subsequent invalid data and ensuring the stability of production data. By linking the entire chain of data with a digital twin and feeding back into the processing rules, a closed-loop optimization of data from processing to application is achieved, continuously improving the system's long-term processing capabilities and meeting the needs of precise quality control in industrial production. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the structure of a zinc-aluminum-magnesium steel strip surface scratch data processing system provided in an embodiment of the present invention;

[0055] Figure 2 This is a main flowchart of a method for processing surface scratch data of zinc-aluminum-magnesium steel strip provided in an embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Please see Figure 1 This invention provides a data processing system for surface scratches on zinc-aluminum-magnesium steel strips, comprising:

[0059] The material parameter input module is used to receive and verify the basic parameters of zinc-aluminum-magnesium steel strip and output the parameter package that has passed the verification.

[0060] The raw scratch data input module is used to receive the raw scratch data block output by the front-end detection device, perform integrity verification on the raw scratch data block, and output the raw data block that has passed the verification.

[0061] A dedicated feature database for zinc, aluminum, and magnesium is used to store dynamic feature parameters associated with the properties of zinc, aluminum, and magnesium steel strips. After receiving the parameter package, the database retrieves and outputs the corresponding feature judgment criteria.

[0062] The multi-source parameter association learning module is used to receive production process parameters and historical training data from the basic parameters and generate dynamic judgment thresholds;

[0063] The dual dynamic verification module receives the original data block that has passed verification, the feature judgment benchmark and the dynamic judgment threshold respectively, performs grayscale verification and morphological verification on the original data block that has passed verification, and outputs a data frame with verification result mark through key judgment logic.

[0064] The invalid data filtering module receives the data frame with the verification result mark, classifies the data frame with the verification result mark, and outputs valid data frames and device abnormal data.

[0065] The bidirectional feedback correction module receives the abnormal data from the equipment, analyzes the root cause of the abnormal data, and outputs equipment adjustment instructions.

[0066] The effective data output and digital twin module receives the effective data frame, outputs a quality report, and constructs a digital twin. The digital twin is used to associate the effective data frame, downstream processing data, and corrosion resistance data, and feeds back into the multi-source parameter association learning module.

[0067] In an optional embodiment, the material parameter input module includes:

[0068] The parameter receiving unit is used to receive the basic parameters of the zinc-aluminum-magnesium steel strip, including batch number, coating thickness, magnesium content, rolling completion time and production equipment number.

[0069] The parameter verification unit is used to verify the format, value range and validity of the basic parameters. If the verification fails, an exception code is output and a preset processing action is executed. The preset processing action includes refusing to receive or triggering manual confirmation.

[0070] The data interface unit uses a preset industrial Ethernet protocol to interface with the production execution system. The transmission rate of the preset industrial Ethernet protocol is not less than 100Mbps, and the parameter transmission delay is not more than 100ms.

[0071] The parameter package output unit is used to encapsulate the verified basic parameters into the parameter package and send it synchronously to the original scratch data input module and the zinc-aluminum-magnesium exclusive feature database.

[0072] The parameter receiving unit is responsible for collecting basic parameters of zinc-aluminum-magnesium steel strip. The production equipment number in the basic parameters refers to the unique identifier of the front-end detection equipment used to collect scratch data for this batch of steel strip (e.g., CCD-001, Laser-002), which is crucial for subsequent equipment anomaly detection. The parameter verification unit performs multi-level checks on the format, value range, and validity of the basic parameters. The "trigger manual confirmation" preset processing action means that when a parameter has a non-fatal anomaly (e.g., magnesium content slightly exceeding the standard range of 1.5%-3.0% but belonging to a specially customized batch), the system will pop up an interactive window prompting the operator to verify and enter confirmation opinions, avoiding direct rejection. Production interruption; the data interface unit refers to the functional unit that enables data interaction between the system and the Production Execution System (MES system). The preset industrial Ethernet protocol refers to a communication protocol (such as Profinet) that meets the data transmission requirements of industrial scenarios. A transmission rate of 100Mbps and a latency of ≤100ms ensure the real-time transmission of parameters, adapting to the production rhythm of 20-30 meters per minute for zinc-aluminum-magnesium steel strip; the parameter package output unit refers to the functional unit that encapsulates the verified basic parameters into standardized data packets and distributes them, synchronously sending them to the original scratch data input module and the zinc-aluminum-magnesium dedicated feature database. This is the key to achieving "strong binding between material parameters and subsequent processing links," avoiding data processing deviations due to parameter transmission delays or omissions. Overall, these four units form a closed loop of "receiving-verification-transmission-output," which not only ensures the accuracy of basic parameters but also improves parameter flow efficiency, laying the foundation for the overall processing accuracy of the system.

[0073] In one possible implementation, the parameter receiving unit supports batch parameter import. When multiple batches of zinc-aluminum-magnesium steel strips need to be processed within a certain time period, the system can receive multiple batches of basic parameters (e.g., 10 batches) at once through the data interface unit and automatically sort them by batch number. The parameter verification unit verifies the imported parameters batch by batch, and the verification results are displayed in a list format (e.g., "ZM20240601001: Pass; ZM20240601002: Magnesium content abnormal (3.1%), awaiting manual confirmation"). Operators can process items awaiting confirmation in batches. The data interface unit also supports data interruption resumption. If the network is interrupted during transmission, the system will automatically resume transmitting the unfinished parameters from the interrupted position after the connection is restored, without needing to re-upload all the data. The parameter packet output unit adds a unique checksum (e.g., MD5 code) to each parameter packet. Subsequent receiving modules can verify whether the parameter packet has been tampered with during transmission using the checksum, further ensuring data integrity.

[0074] Additionally, it should be noted that the "validity verification" in the parameter verification unit not only includes format and value range checks, but also performs logical verification by referencing historical data. For example, when verifying the production equipment number, the system will query the equipment's historical maintenance records. If the equipment is in a "fault repair" state, the production equipment number will be deemed invalid and an exception code P005 will be output to avoid using parameters from faulty equipment that could lead to deviations in subsequent data collection. After the parameter package output unit distributes the parameter packages, it will record the reception status of each parameter package (e.g., "Original scratch data input module has received; Zinc-aluminum-magnesium specific feature database has received"). If a receiving module fails to report successful reception within 5 seconds, the system will resend the parameter package, retrying up to 3 times. If it still fails, an alarm will be triggered to ensure that the parameter package is successfully received by all required modules.

[0075] For example, in one feasible implementation, the MES system of a zinc-aluminum-magnesium steel strip production workshop stores the basic parameters of three batches of steel strips: batch 1 (ZM20240601001, coating 8.0μm, magnesium content 2.2%, rolling time 2024-06-01 08:00:00, equipment CCD-001), batch 2 (ZM20240601002, coating 9.5μm, magnesium content 2.8%, rolling time 2024-06-01 08:15:00, equipment CCD-002), and batch 3 (ZM20240601003, coating 15.5μm, magnesium content 2.5%, rolling time 2024-06-01 08:30:00, equipment Laser-001). The parameter receiving unit of the material parameter input module receives these three batches of parameters from the MES system in batches via the Profinet protocol. The parameter verification unit verifies each batch: Batch 1's format and values ​​are compliant and pass; Batch 2's magnesium content of 2.8% is within the range of 1.5%-3.0%, and the equipment's CCD-002 historical maintenance record is normal and pass; Batch 3's coating thickness of 15.5μm exceeds the range of 5.0-15.0μm, outputting an error code P002 and prompting manual confirmation. After verification by the operator, it was found that the error was entered into the MES system (actual thickness was 14.5μm), and after correction, the verification passed; The data interface unit maintains a transmission rate of 100Mbps throughout the process, and the parameter transmission delay is ≤50ms; The parameter package output unit encapsulates the three batches of verified parameters into parameter packages with MD5 checksums and sends them synchronously to the original scratch data input module and the zinc-aluminum-magnesium exclusive feature database. After receiving successful reception feedback from both modules, it records "The parameter packages of the three batches have been successfully distributed" in the system log.

[0076] In an optional embodiment, the parameter verification unit includes:

[0077] The format verification subunit is used to verify whether the format of the batch number is a preset identifier + 6-digit date + 3-digit serial number, wherein the preset identifier is "ZM" and the total length of the batch number is fixed at 12 digits;

[0078] The value verification subunit is used to verify whether the value range of the coating thickness is 5.0-15.0μm, whether the value range of the magnesium content is 1.5%~3.0%, whether the rolling completion time is not a future time, and whether the production equipment number is in the preset equipment list.

[0079] An exception handling subunit is used to output a corresponding exception code when the basic parameters do not meet the verification requirements of the format verification subunit or the value verification subunit. The exception codes include P001 for batch number format exception, P002 for coating thickness exception, P003 for magnesium content exception, P004 for rolling completion time exception, and P005 for production equipment number exception.

[0080] The format verification subunit is a functional subunit specifically designed to check the format standardization of basic parameters. The preset identifier "ZM" refers to the exclusive prefix of the batch number of zinc-aluminum-magnesium steel strip (distinguished from "DX" for galvanized sheet and "BX" for stainless steel). The fixed length of 12 digits ("ZM" + 6 digits for date + 3 digits for serial number) is to ensure the uniqueness and traceability of the batch number. For example, in "ZM20240601001", "20240601" represents June 1, 2024, and "001" represents the first batch on that day, facilitating quick location of batch information. The value verification subunit is a functional subunit that checks whether the values ​​of basic parameters are within the reasonable range of zinc-aluminum-magnesium steel strip. The coating thickness range of 5.0-15.0μm and magnesium content range of 1.5%-3.0% are based on the industrial production standards of zinc-aluminum-magnesium steel strip (such as the coating thickness of zinc-aluminum-magnesium steel strip used in automobiles is usually...). The rolling completion time (not a future time) is determined by the material characteristics (excessive magnesium content can easily lead to increased coating brittleness) and is set to ensure the authenticity of the parameters. The preset equipment list refers to the list of front-end testing equipment that has been registered and is operating normally in the system (such as CCD-001, Laser-001), avoiding the inclusion of invalid equipment numbers. The anomaly handling subunit is a functional subunit that performs standardized anomaly handling for parameters that fail format or value verification. Anomaly codes P001-P005 are used to clearly distinguish different types of anomalies, making it easier for operators to quickly locate problems (e.g., seeing P002 indicates an abnormal coating thickness). At the same time, the system will provide a solution prompt for each anomaly code (e.g., the prompt for P003 is "Magnesium content must be within the range of 1.5%-3.0%; if it is a special customized batch, please confirm manually"), improving the efficiency of anomaly handling. Overall, these three subunits construct a comprehensive parameter verification system from the three dimensions of "format-value-anomaly," which not only ensures the compliance of parameters but also provides clear guidance for anomaly handling, and is the core guarantee for the accuracy of the material parameter input module.

[0081] In one possible implementation, the format verification subunit supports custom batch number formats. If a company has special batch number coding rules (such as "ZM" + 4-digit year + 2-digit month + 2-digit date + 4-digit serial number), operators can modify the format rules in the system settings, and the system will verify according to the new rules. The value range of the value verification subunit can be adjusted according to the product type. For example, for zinc-aluminum-magnesium steel strips used in home appliances, the coating thickness range can be adjusted to 5.0-10.0μm (home appliances have slightly lower corrosion resistance requirements than automobiles), and for steel strips used in pipelines, it can be adjusted to 10μm. For pipes with a diameter of 0-15.0μm (higher corrosion resistance is required), the system stores multiple value range templates, which operators can quickly select according to the production order type. The exception handling subunit supports batch export of exception codes. When multiple exceptions occur in the verification of a batch of parameters (such as the simultaneous existence of batch number format exceptions and production equipment number exceptions), the system will generate an Excel report containing the exception code, corresponding parameters, and exception reasons, which is convenient for operators to archive and analyze exception trends (such as frequent P004 exceptions within a certain period of time, which may be due to MES system time synchronization failure).

[0082] Additionally, it should be noted that the format verification subunit's batch number verification also includes "uniqueness verification." The system will query historical parameter records; if the currently entered batch number already exists (e.g., duplicate entry of ZM20240601001), the format verification subunit will determine it as a format anomaly and output P001 to prevent duplicate processing of data from the same batch. The value verification subunit's verification of the rolling completion time will also calculate the storage time based on the current time. If the storage time exceeds 168 hours (7 days), the system will display a prompt after the value verification passes: "This batch has been stored for a long time; it is recommended to pay close attention to oxidation and false scratches," providing an early warning for subsequent data processing. The anomaly handling subunit will initiate anomaly tracking after outputting the anomaly code. If the operator does not handle the anomaly within 30 minutes, the system will send a text message or email reminder to production management personnel to prevent anomaly accumulation from affecting production progress.

[0083] For example, in one feasible implementation, an operator inputs the following basic parameters for a batch of zinc-aluminum-magnesium steel strip into the parameter verification unit: batch number "ZM20240601001" (duplicate entry), coating thickness 16.0μm, magnesium content 2.3%, rolling completion time 2024-06-10 10:00:00 (future time), and production equipment number "CCD-005" (not in the preset list). The format verification subunit first verifies the batch number and finds that it already exists in the history, outputting exception code P001; the value verification subunit verifies that the coating thickness of 16.0μm exceeds the range of 5.0-15.0μm, outputting P002; the rolling completion time is a future time, outputting P004; and the production equipment number "CCD-005" is not in the preset list, outputting P005; the exception handling subunit integrates these exceptions and displays "Exception codes: P001 (duplicate batch number), P002 (coating thickness out of range), P005 (duplicate batch number), P002 (coating thickness out of range), P003 (duplicate batch number), P004 (duplicate batch number), P005 (duplicate batch number), P002 (duplicate batch number ...3 (duplicate batch number), P004 (duplicate batch number), P005 (duplicate batch number), P003 (duplicate batch number), The system returned "04 (rolling time is in the future)" and "P005 (invalid equipment number)," and provided solution suggestions for each code. After the operator corrected the parameters to batch number "ZM20240601002", coating thickness 14.5μm, rolling completion time 2024-06-01 10:00:00, and production equipment number "CCD-002", the system re-verified the parameters. The format verification subunit confirmed that the batch number was unique and the format was correct, the value verification subunit confirmed that all parameters were within a reasonable range, and the exception handling subunit did not output any exception codes. The parameter verification passed.

[0084] In an optional embodiment, the raw scratch data input module includes:

[0085] The data block receiving unit is used to receive the original scratch data block sent by the front-end detection device. The original scratch data block is encapsulated in JSON format. Every 100 frames of data constitute one original scratch data block. The header of the original scratch data block contains a batch number, a data block number, and a verification field for the number of frames.

[0086] The integrity verification unit is used to verify whether the actual number of frames in the original scratch data block is consistent with the number of frames in the header, whether the detection area coordinates in the original scratch data block are not repeated, and whether the scratch depth and scratch length in the original scratch data block are non-negative values.

[0087] The data block processing unit is used to output the original data block that has passed the integrity verification unit to the dual dynamic verification module when the original scratch data block passes the verification; when the original scratch data block fails the verification, it triggers a retransmission instruction, the maximum number of times the retransmission instruction is triggered is 3. If the verification still fails after 3 retransmissions, the original scratch data block is marked as an abnormal data block and sent to the invalid data filtering module.

[0088] The data block receiving unit is a functional unit that receives data blocks encapsulated by the front-end detection device. JSON format is a lightweight data exchange format that is easy for machines to parse and generate. Each data block consists of 100 frames of data, designed to balance data transmission efficiency and processing pressure. This avoids frequent network interactions caused by single-frame transmission and prevents processing delays due to excessively large data blocks. The verification fields in the data block header (batch number, data block number, frame count) are used for quick data attribution and integrity checks. For example, the batch number can be used to associate a data block with the corresponding material parameters, and the data block number can be used to track the transmission order of the data blocks. The integrity verification unit checks whether there are missing data or logical contradictions in the data block. The functional units ensure data block integrity by maintaining consistent frame counts to prevent frame loss or redundancy, ensuring no duplicate detection area coordinates to prevent repeated processing of scratch data at the same location, and guaranteeing the physical validity of the data by ensuring non-negative scratch depth and length (scratch depth / length cannot be negative). The data block processing unit distributes data blocks based on verification results. The retransmission instruction is a command sent by the system to the front-end detection equipment requesting the retransmission of abnormal data blocks. The three-retransmission limit is based on a balance between data integrity and production efficiency, avoiding production interruptions caused by unlimited retransmissions. Abnormal data blocks are those that fail verification after multiple retransmissions and are stored in an abnormal data cache for subsequent manual investigation (e.g., analyzing whether it's a hardware failure). Overall, these three units form a "receive-verify-process" chain, ensuring both the quality of the original data and production continuity through a reasonable anomaly handling mechanism, forming the foundation for the system's accurate identification of valid scratches.

[0089] In one possible implementation, the data block receiving unit supports parallel reception from multiple devices. That is, the system can simultaneously receive data blocks output from multiple front-end detection devices (such as two CCD cameras and one laser profilometer). The data source is distinguished by device number, and the data blocks are named according to the rule of "device number + data block number" (e.g., "CCD-001_001") to avoid data block confusion. The integrity verification unit also verifies the field integrity of the data frame, checking whether each frame contains all required fields: "detection area coordinates, grayscale value, scratch depth, scratch length, and data validity marker." If a frame is missing... If the "scratch depth" field is missing, the data block is determined to have failed the verification. When marking abnormal data blocks, the data block processing unit records detailed abnormal information (such as "Data block 001: Frame count mismatch (actual 98 frames ≠ header 100 frames), 3 retransmissions failed") and associates it with the corresponding front-end detection device number, which helps maintenance personnel quickly locate faulty devices. In addition, the data block processing unit also supports manual triggering of retransmission for abnormal data blocks. If the cause of the abnormality is confirmed to be resolved after manual investigation (such as re-plugging and unplugging the device interface), the corresponding abnormal data block can be selected in the system and a retransmission command can be sent manually without waiting for the system to process it automatically.

[0090] Additionally, it should be noted that the data block receiving unit monitors network bandwidth usage in real time when receiving data blocks. If the bandwidth usage exceeds 80% (e.g., multiple devices simultaneously transmitting large data blocks), the system will automatically adjust the data block receiving priority, prioritizing the receiving of data blocks corresponding to the currently producing batch and temporarily suspending the receiving of historical batch data blocks that have completed production, ensuring that real-time production data processing is not affected. The integrity verification unit's verification speed can be adapted to the acquisition speed of the front-end detection device. For example, if the front-end device outputs one data block per second (100 frames), the integrity verification unit can complete the verification of at least two data blocks per second, avoiding data block accumulation. When the data block processing unit outputs the original data blocks that have passed verification, it will add a timestamp (accurate to milliseconds) to each data block. The subsequent dual dynamic verification module can use the timestamp to determine the data processing order, ensuring that data blocks are processed in the order of the production process, avoiding scratch position matching errors caused by data out of order.

[0091] For example, in one feasible implementation, the front-end detection device CCD-001 (acquiring 1 data block per second) and Laser-001 (acquiring 0.5 data blocks per second) simultaneously transmit data to the system. The data block receiving unit receives data from both devices in parallel, naming the data blocks of CCD-001 as "CCD-001_001" and "CCD-001_002", and the data blocks of Laser-001 as "Laser-001_001". When the network bandwidth utilization reaches 75%, it prioritizes receiving the CCD-001 data block corresponding to the current production batch (ZM20240601001). After receiving "CCD-001_001", the integrity verification unit first verifies the number of frames (actual 100 frames are consistent with the 100 frames in the header), then verifies the coordinates of the detection area (no duplicates), and finally verifies the scratch depth / length (both are non-negative values), and determines that it passes the test. The unit then receives the data from Laser-001. When "CCD-001_001" is received, a frame is found to be missing the "scratch length" field, and the verification fails. The data block processing unit sends a retransmission command to the front end. The field is still missing after the second retransmission, but it is complete after the third retransmission, and the verification passes. When "CCD-001_002" is received, there are only 95 frames. After 3 retransmissions, there are still 95 frames. The data block processing unit marks it as an abnormal data block, records "CCD-001_002: Frame count mismatch, 3 retransmissions failed", and pushes it to the abnormal data buffer. Finally, the data block processing unit takes "CCD-001_001" and "Laser-001_001" as the original data blocks that have passed the verification, adds a millisecond-level timestamp, and outputs them to the dual dynamic verification module.

[0092] In an optional embodiment, the zinc-aluminum-magnesium specific feature database includes:

[0093] A three-dimensional index storage unit is used to store a tree-like index structure with coating thickness as the first-level index, magnesium content as the second-level index, and storage time as the third-level index. The storage time is divided into intervals of 0-24h, 24-48h, 48-72h, and greater than 72h.

[0094] The oxidation model unit has a built-in oxide layer thickness calculation model, and its calculation formula is as follows: ,in, This indicates the thickness of the oxide layer, in μm, with a value ranging from 0 to 2.0 μm. The rate constant is expressed in units of 1000 ppm. The correlation between the magnesium content and the magnesium content is positive, specifically: when the magnesium content is 1.5%, =0.02; when the magnesium content is 2.0%, =0.025; when the magnesium content is 2.5%, =0.03; when the magnesium content is 3.0%, =0.04; This represents the storage time, in hours, and ranges from 0 to 168 hours.

[0095] The feature benchmark output unit is used to receive the parameter package, calculate the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieve the three-dimensional index storage unit and call the oxidation model unit to correct the grayscale range, and output the feature judgment benchmark containing the grayscale range, grayscale compensation coefficient and depth-length rule to the dual dynamic verification module.

[0096] The three-dimensional index storage unit refers to a functional unit that stores dynamic feature parameters according to three dimensions: "coating thickness - magnesium content - storage time". The tree-shaped index structure refers to a hierarchical storage structure similar to "trunk - branches - leaves". The first-level index, coating thickness, is divided in 1.0μm increments (5.0μm, 6.0μm...15.0μm); the second-level index, magnesium content, is divided in 0.5% increments (1.5%, 2.0%...3.0%); and the third-level index, storage time, is divided by time period (0-24h, 24-48h, etc.). This division is based on experimental data of zinc-aluminum-magnesium materials (such as the grayscale range variation under different coating thicknesses) to ensure that the feature parameters corresponding to the index can accurately match the actual material characteristics. The oxidation model unit refers to a functional unit with a built-in oxide layer thickness calculation model. The oxide layer thickness calculation formula... It is a zinc-aluminum-magnesium specific oxidation kinetic model derived from the Arrhenius equation, in which the rate constant... Positively correlated with magnesium content (higher magnesium content, faster oxidation rate), value range (0.02-0.04 μm / The oxide layer thickness was determined by fitting data from 100,000 sets of zinc, aluminum, and magnesium oxidation experiments. The 0-2.0μm range corresponds to the oxidation level of zinc-aluminum-magnesium steel strip within the normal storage time (0-168h). Exceeding this range requires a reassessment of material properties. The feature reference output unit refers to the functional unit that retrieves and corrects feature parameters based on material parameters and then outputs a judgment reference. Calculating the storage time involves subtracting the rolling completion time from the current time (e.g., subtracting 8:00 from the current 10:30 to get 2.5h). Calling the oxidation model to correct the grayscale range involves adjusting the base grayscale range based on the calculated oxide layer thickness (e.g., increasing the oxide layer thickness by 0.04μm by 1, raising the upper limit of the grayscale range by 1), ensuring that the output grayscale range, grayscale compensation coefficient, and depth-length rules are adapted to the actual oxidation state and material properties of the current batch of steel strip. Overall, these three units realize the functions of "feature parameter storage - oxidation correction - reference output," ensuring both the specific adaptability of feature parameters and dynamic adjustment to cope with changes in material state, which is the core guarantee for the accuracy of dual dynamic verification.

[0097] In one possible implementation, the index step size of the three-dimensional index storage unit can be refined according to actual production needs. For example, the coating thickness step size can be adjusted from 1.0 μm to 0.5 μm (5.0 μm, 5.5 μm...15.0 μm), the magnesium content step size can be adjusted from 0.5% to 0.2% (1.5%, 1.7%...3.0%), and the storage time intervals can be adjusted from 4 to 6 (0-12h, 12-24h...120-168h). A more refined index division can further improve the adaptability of feature parameters. For example, for a special batch with a coating thickness of 5.5 μm, a magnesium content of 1.7%, and a storage time of 18h, a more accurate grayscale interval can be retrieved. The oxidation model unit can also introduce a production temperature correction factor, that is, adjust the rate constant according to the production temperature in the material parameters. For example, when the production temperature is 530℃ (higher than the conventional 510℃), The value is increased by 10% from the original value (e.g., when the magnesium content is 2.2%). The value was adjusted from 0.025 to 0.0275 to better reflect the actual impact of production temperature on the oxidation rate. When outputting the feature judgment benchmark, the feature benchmark output unit will also output the corresponding index matching record (such as "matching index: coating 8.0μm + magnesium content 2.2% + storage 0-24h, oxide layer thickness 0.04μm, gray range correction +1"), which facilitates the subsequent traceability of the benchmark source. If a verification deviation occurs, it can be quickly located whether the problem is caused by index matching or oxidation correction.

[0098] Additionally, it should be noted that the 3D index storage unit supports dynamic updates. Every 5000 sets of actual zinc-aluminum-magnesium scratch data (including manually labeled valid scratches and pseudo-scratches) accumulated, the system automatically triggers index optimization. This optimization uses a linear regression algorithm to correct the grayscale range and depth-length rules corresponding to each level of the index. For example, if, under a certain index combination (8.0μm coating + 2.2% magnesium content + 0-24h storage), the grayscale values ​​of actual pseudo-scratches are mostly concentrated in the 175-176 range, the system will lower the upper limit of the grayscale range for that index from 176 to 174, reducing false scratch misjudgments. The oxidation model unit... The values ​​will be refitted periodically (e.g., every 3 months) based on newly accumulated oxidation experimental data to ensure... The value reflects the latest material oxidation characteristics; if the feature reference output unit cannot find a completely matching three-dimensional index (e.g., magnesium content 2.3%, no secondary index for 2.3%), it will use interpolation to calculate and generate feature parameters, for example, based on a magnesium content of 2.2% ( =0.025) and 2.5% ( The parameter (=0.03) is used to calculate 2.3% proportionally. The value (0.025+(2.3-2.2) / (2.5-2.2)×(0.03-0.025)=0.0267) ensures that accurate feature determination benchmarks can still be output even when there are no missing indexes.

[0099] For example, in one feasible implementation, the zinc-aluminum-magnesium specific feature database receives the parameter package output by the material parameter input module: batch number ZM20240601001, coating thickness 8.0μm, magnesium content 2.2%, rolling completion time 2024-06-01 08:00:00, and production temperature 520℃. The three-dimensional index storage unit first uses 8.0μm as the primary index, 2.2% as the secondary index (no exact match, closest values ​​are 2.0% and 2.5%), and the storage time of 2.5h (0-24h interval) calculated at the current time 10:30 as the tertiary index to determine the basic index combination; the oxidation model unit calls the formula. Because the production temperature of 520℃ is higher than the conventional 510℃, the magnesium content of 2.2% corresponds to... The value was increased by 10% from 0.025 to 0.0275, and the calculation yielded... =0.0275×(2.5 The feature benchmark output unit interpolates the base grayscale range of 2.2% to 136-176 based on the basic index combination (8.0μm+2.0%+0-24h: grayscale 135-175; 8.0μm+2.5%+0-24h: grayscale 137-177). Then, based on the oxide layer thickness of 0.0437μm, the upper limit of the grayscale range is increased by 1 to 136-177. At the same time, the grayscale compensation coefficient "strong light -3, weak light +3" and the depth-length rule "depth ≥ 0.03μm → length ≥ 0.2mm" are determined. Finally, the feature benchmark output unit outputs the feature judgment benchmark containing "grayscale range 136-177, grayscale compensation coefficient, depth-length rule, index matching record, and oxide correction record" to the dual dynamic verification module.

[0100] In an optional embodiment, the dual dynamic verification module includes:

[0101] A data integration unit is used to integrate the verified original data block, the feature determination benchmark, and the dynamic determination threshold.

[0102] The grayscale verification unit is used to calculate the storage time based on the collection timestamp in the verified original data block, call the oxidation model unit to calculate the oxide layer thickness to correct the grayscale range, and combine the light intensity to apply the grayscale compensation coefficient to determine whether the grayscale value in the verified original data block is within the corrected grayscale range.

[0103] The morphological verification unit is used to extract the scratch depth and scratch length from the original data block that has passed the verification, calculate the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and determine whether the ratio is not less than 1.

[0104] The key judgment unit is used to execute the key judgment logic to determine whether the original data block that has passed the verification has passed the verification of the grayscale verification unit and the morphological verification unit at the same time. If it passes at the same time, it is marked as data to be confirmed as valid data. If it does not pass at the same time, it is marked as invalid data to be classified. The unit outputs the data frame with the verification result mark containing the data to be confirmed as valid data and the data to be classified as invalid data.

[0105] The data integration unit is a functional unit that integrates three types of data: raw data, feature benchmarks, and dynamic thresholds. The purpose of integration is to ensure that subsequent verification processes can directly access the required data, avoiding verification delays or omissions caused by data dispersion. For example, each frame of data in a verified raw data block is associated with its corresponding feature judgment benchmark (e.g., the grayscale range corresponding to the data acquisition time) and dynamic judgment threshold (e.g., the minimum length of the production process corresponding to the data frame), forming an independent verification unit of "one frame of data + corresponding benchmark + corresponding threshold." The grayscale verification unit is a functional unit that verifies the validity of data based on the specific grayscale features of zinc, aluminum, and magnesium. The acquisition timestamp refers to the time (accurate to milliseconds) recorded when the front-end detection equipment acquires the data frame, used to calculate the storage time corresponding to the data frame (e.g., acquisition time 10:25 minus rolling time 8:00 equals 2.42h). Calling the oxidation model to correct the grayscale range refers to fine-tuning the basic grayscale range based on the oxide layer thickness calculated according to the storage time (e.g., if the oxide layer thickness increases by 0.04μm, the upper limit of the range is increased by 1). The light intensity refers to the environmental intensity simultaneously acquired by the front-end detection equipment. Illumination data (0-1000 lux) uses a grayscale compensation coefficient (e.g., no compensation is needed for strong light at 650 lux) to ensure that grayscale verification is not affected by changes in illumination. The morphology verification unit is a functional unit that verifies the validity of data based on scratch morphology characteristics (depth-length). Extracting scratch depth and length means reading the corresponding values ​​from the original data frame (e.g., depth 0.05μm, length 0.22mm). Calculating the ratio means dividing the actual scratch length by the minimum length in the dynamic judgment threshold (e.g., 0.22 / 0.19≈1.158). A ratio ≥1 indicates that the morphology meets the requirements for valid scratches. This judgment logic is based on the quality standards of zinc-aluminum-magnesium steel strips (e.g., for scratches with a depth greater than 0.03μm, the length must be ≥0.2mm to affect corrosion resistance). The key judgment unit is a functional unit that executes the key judgment logic. The key judgment logic is the rule that "data is judged as valid only after passing both grayscale verification and morphology verification." Marking data to be confirmed as valid and data to be classified as invalid is to provide a clear classification basis for the subsequent invalid data filtering module and avoid classification errors caused by data mixing. Overall, these four units form a verification chain of "data preparation - grayscale screening - morphology screening - result labeling", which fully combines the characteristics of zinc, aluminum and magnesium materials (grayscale verification) and adapts to the production process conditions (morphology verification), and is the core link for the system to accurately identify effective scratches.

[0106] In one possible implementation, the data integration unit supports data anomaly warning. If a certain type of data is found to be missing during the integration process (such as the lack of a feature determination benchmark for the corresponding frame data), a warning message will be immediately output (such as "Frame X100Y200: Feature determination benchmark missing"), and the verification of that frame data will be paused until the data is supplemented before continuing, thus avoiding misjudgment caused by the lack of benchmark verification. The illumination compensation rules of the grayscale verification unit can be further refined, for example, dividing the illumination intensity into four intervals: weak light (<300 lux), normal light (300-800 lux), strong light (>800 lux), and extremely strong light (>1200 lux), with the grayscale area of ​​the extremely strong light interval being... The upper and lower limits are each lowered by 5 to more accurately cope with extreme lighting environments; if the morphology verification unit finds slight fluctuations in scratch depth data (such as a depth of 0.031μm in one frame and 0.030μm and 0.029μm in two adjacent frames), it will use the average depth of three frames (0.030μm) for calculation to reduce the impact of data acquisition errors on verification results; in addition to marking valid data to be confirmed and invalid data to be classified, the key judgment unit will also mark data that "passes only one verification" as "to be reviewed" (such as marking "morphology to be reviewed" if only grayscale verification is passed) and temporarily store it in the temporary cache area to support manual review and avoid missing valid data due to single verification errors.

[0107] Additionally, it should be noted that the data integration unit adds a unique identifier (such as "batch number + frame coordinates") to the integrated data to facilitate subsequent traceability and verification results. If there is a dispute regarding the verification result of a certain frame of data, the original data, feature baseline, and dynamic threshold of that frame can be quickly located through the unique identifier. The grayscale verification unit records the grayscale verification process of each frame of data (e.g., "Frame X100Y200: storage time 2.42h → oxide layer 0.04μm → grayscale range 136-177, illumination 650lux → no compensation, grayscale 155 in..."). Within the range → Pass), the morphological verification unit will also record the morphological verification process (e.g., "Frame X100Y200: Depth 0.05μm≥0.03μm, Length 0.22mm / 0.19mm=1.158≥1 → Pass"). These records will be output along with the verification results to facilitate subsequent analysis of the reasons for verification deviations. When the key judgment unit outputs data frames marked with verification results, it will sort them by batch number and acquisition timestamp to ensure that the order of data frames is consistent with the production process and to avoid data disorder during subsequent processing by the invalid data filtering module.

[0108] For example, in one feasible implementation, the dual dynamic verification module receives the verified original data block (frame X100.2Y200.5: grayscale 155, depth 0.05μm, length 0.22mm, acquisition time 10:25:00), feature judgment criteria (grayscale 136-177, strong light -3, depth ≥0.03μm → length ≥0.2mm), and dynamic judgment threshold (depth ≥0.03μm → minimum length 0.19mm). The data integration unit associates the frame data with feature benchmarks and dynamic thresholds to generate an independent verification unit and adds a unique identifier "ZM20240601001_X100.2Y200.5"; the grayscale verification unit calculates the storage time as 2.42h based on the acquisition time of 10:25, calls the oxidation model to confirm that the oxide layer thickness is 0.04μm, and no additional correction is needed for the grayscale range. The light intensity received from the front-end device is 650 lux (normal light), which does not require compensation. It determines that the grayscale value of 155 is within the range of 136-177, records the verification process, and determines that it has passed; the morphological verification unit extracts a depth of 0. The frame has a grayscale value of 0.05μm and a length of 0.22mm. The calculated ratio is 0.22 / 0.19≈1.158≥1. The verification process is recorded and the frame is deemed to have passed. The key judgment unit executes the logic to confirm that the frame has passed both verifications. It is marked as "valid data to be confirmed" and the data frame with the verification result mark, which includes a unique identifier, verification process record, and verification result mark, is output to the invalid data filtering module. At the same time, another frame of data (grayscale 178, depth 0.003μm, length 0.2mm) is verified. The grayscale 178 exceeds the range of 136-177 and is marked as "invalid data to be classified". It is also output together.

[0109] In an optional embodiment, the effective data output and digital twin module includes:

[0110] The quality report generation unit is used to compile the valid data frames into a batch summary table and a list of valid scratch coordinates. The batch summary table includes the batch number, the total number of valid scratches, the number of scratches in different depth ranges, the maximum scratch size, and the data pass rate. The list of valid scratch coordinates includes the detection area coordinates, scratch depth, scratch length, and collection timestamp.

[0111] A digital twin construction unit is used to associate the batch summary table, downstream processing data, and corrosion resistance data to construct the digital twin. The downstream processing data includes scratch extension records in the stamping process, and the corrosion resistance data includes salt spray test results.

[0112] The data feedback unit is used to send the data in the digital twin to the multi-source parameter association learning module to trigger incremental training of the multi-source parameter association learning module and update the dynamic judgment threshold.

[0113] The quality report generation unit refers to the functional unit that compiles valid data frames into a standardized quality report. The batch summary table is a table that compiles key quality indicators for scratches by batch. The total number of valid scratches refers to the number of scratch data frames that have passed all verifications in the batch. The number of scratches in different depth ranges (e.g., 0.03-0.05μm, >0.05μm) is used to distinguish the degree of impact of scratches on quality (the greater the depth, the more serious the impact). The maximum scratch size refers to the maximum combination of scratch depth and length in the batch (e.g., 0.08μm × 2.5mm). The data pass rate refers to the ratio of the number of valid data frames to the total number of data frames (e.g., 92 / 100 = 92%). These indicators directly reflect the batch quality level and are the core basis for factory quality inspection and customer acceptance. The valid scratch coordinate list refers to a list that records the specific location and characteristics of each valid data frame, which is convenient for subsequent traceability (e.g., if the customer reports a scratch at a certain location). When troubleshooting, the corresponding data can be quickly located. The digital twin building unit refers to the functional unit that builds a virtual model that corresponds one-to-one with the physical batch. The digital twin refers to the virtual model that integrates production data (quality report), downstream processing data (such as whether scratches extend in the stamping process), and corrosion resistance data (such as whether corrosion occurs after 500 hours of salt spray testing). Its purpose is to achieve full lifecycle traceability "from production to application" and verify the actual value of scratch data processing results. The data feedback unit refers to the functional unit that uses digital twin data to optimize system rules. The core of feedback is to transform actual application feedback (such as no corrosion in salt spray testing for scratches of a certain depth) into the basis for adjusting dynamic judgment thresholds (such as adjusting the minimum length threshold), so as to achieve system self-optimization. From an overall perspective, these three units not only meet the quality control needs of the current batch, but also provide data support for the long-term accuracy improvement of the system, which is the key to maximizing the value of data.

[0114] In one possible implementation, the quality report generation unit supports custom report templates, allowing enterprises to adjust the fields of the batch summary table according to their needs (such as adding "number of scratches in different length ranges" and "statistics on scratch distribution areas"). The list of effective scratch coordinates supports map-style display (marking scratch locations on a two-dimensional planar map of the steel strip and using colors to distinguish depth ranges), providing a more intuitive presentation of scratch distribution. The digital twin construction unit can link more dimensional data, such as raw material composition data (specific contents of zinc, magnesium, and aluminum) and rolling process parameters (rolling speed, annealing temperature), analyzing the correlation between these factors and scratch quality (e.g., whether excessive magnesium content leads to an increase in scratches). The data feedback unit's feedback logic can be set with trigger conditions. For example, if the corrosion rate of a scratch corresponding to a certain dynamic judgment threshold is less than 5% in the salt spray test, the threshold adjustment is triggered (such as lowering the minimum length threshold). If the corrosion rate is higher than 20%, the threshold tightening is triggered (such as raising the minimum length threshold) to ensure that the feedback direction meets the quality requirements. In addition, the quality report generation unit also supports automatic report push. After generation, the report is immediately sent to the email or system account of production management personnel and quality inspectors. The data feedback unit will record the detailed information of each feedback (such as "2024-06-01: Based on the salt spray data of batch ZM20240601001, the minimum length threshold for depth ≥0.03μm is lowered from 0.19mm to 0.18mm"), which is convenient for tracing the reason for the rule adjustment.

[0115] Additionally, it should be noted that the PDF reports generated by the quality report generation unit automatically add electronic signatures and anti-counterfeiting codes. The anti-counterfeiting codes can be verified through the company's official website to prevent report tampering. The list of valid scratch coordinates can be exported to CAD format, making it easier for downstream customers to avoid scratched areas during processing. The digital twin construction unit supports virtual simulation, which can simulate the impact of different processing techniques (such as stamping and bending) on ​​scratches and predict whether scratches will expand into quality defects after processing. The incremental training of the data feedback unit supports manual confirmation. If the threshold adjustment suggested by the feedback is large (such as decreasing from 0.19mm to 0.15mm), the system will prompt technical personnel to review it. The adjustment will be implemented only after confirmation to avoid rule deviations caused by abnormal data. Furthermore, the digital twin supports historical backtracking, allowing querying of digital twin data from any historical batch (such as quality reports from a batch one year ago, downstream processing records) for quality trend analysis (such as whether the scratch pass rate has improved in a certain quarter).

[0116] For example, in one feasible implementation, the valid data output and digital twin module receive 92 valid data frames (batch ZM20240601001) output by the invalid data filtering module. The quality report generation unit statistically analyzes the following: total number of valid scratches 92, 68 frames of scratches with a depth of 0.03-0.05μm, 24 frames of scratches >0.05μm, maximum scratch size 0.08μm×2.5mm, and data pass rate 92%. It generates an Excel-format batch summary table and a CSV-format list of valid scratch coordinates (including X / Y coordinates, depth, length, and acquisition time), and simultaneously exports a PDF-format report (including electronic signature and anti-counterfeiting code), which is automatically pushed to the quality inspector's account. The digital twin construction unit associates the production process parameters of this batch (rolling 700m / min, annealing 510℃), stamping data from downstream customers (no scratch extension), and salt spray test data (500h no corrosion). (Corrosion) A digital twin is constructed, and all valid scratch locations are marked on the virtual model. Scratches with a depth of 0.03-0.05μm are marked in blue, and scratches with a depth greater than 0.05μm are marked in red. The data feedback unit analyzes the digital twin data and finds that scratches with a depth of 0.03-0.05μm and a length of 0.18-0.19mm do not corrode in the salt spray test. This triggers incremental training of the multi-source parameter association learning module, which lowers the minimum length threshold for depth ≥ 0.03μm in the dynamic judgment threshold from 0.19mm to 0.18mm, and records the feedback process. Finally, the quality report is pushed to the production execution system, updating the quality status of the batch to "qualified", and the digital twin data is archived to the historical database.

[0117] In this embodiment, by constructing a parameter verification and data processing system specifically for zinc, aluminum, and magnesium, false scratches and invalid data are accurately filtered out, avoiding the problem of poor adaptability of general solutions, thus improving the accuracy of scratch data processing. Through bidirectional feedback between data processing and front-end devices, the quality of raw data collection is optimized from the source, reducing the generation of subsequent invalid data and ensuring the stability of production data. By linking the entire chain of data with a digital twin and feeding back into the processing rules, a closed-loop optimization of data from processing to application is achieved, continuously improving the system's long-term processing capabilities and meeting the needs of precise quality control in industrial production.

[0118] Based on the above embodiments, such as Figure 2 As shown, the present invention also provides a method for processing surface scratch data of zinc-aluminum-magnesium steel strip, comprising the following steps:

[0119] S1: Material parameters and raw data are input synchronously; receive and verify the basic parameters of zinc-aluminum-magnesium steel strip; output the parameter package that has passed verification; receive the raw scratch data block output by the front-end detection equipment; after verifying the integrity of the raw scratch data block, output the raw data block that has passed verification.

[0120] S2: Dedicated feature parameter retrieval, receiving the parameter package, calculating the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieving preset dynamic feature parameters and correcting the grayscale range, and outputting the corresponding feature judgment benchmark;

[0121] S3: Multi-source parameter association rule generation: Receive production process parameters and historical training data from the basic parameters, train the production process parameters and historical training data using a preset algorithm, and generate a dynamic judgment threshold.

[0122] S4: Dual dynamic verification and key judgment, integrating the verified original data block, the feature judgment benchmark, and the dynamic judgment threshold, sequentially performing grayscale verification and morphological verification on the verified original data block, and outputting a data frame marked with the verification result through key judgment logic; wherein, the grayscale verification includes calculating the storage time based on the acquisition timestamp in the verified original data block, correcting the grayscale range by calling the oxide layer thickness calculation formula, applying a grayscale compensation coefficient in combination with the illumination intensity, and judging whether the grayscale value in the verified original data block is within the corrected grayscale range; the morphological verification includes extracting the scratch depth and scratch length in the verified original data block, calculating the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and judging whether the ratio is not less than 1;

[0123] S5: Invalid data filtering and classification: Receive the data frame with verification result mark, classify the data frame with verification result mark according to the preset classification rule, and output valid data frame and device abnormal data;

[0124] S6: Two-way feedback correction: Receive the abnormal data of the device, analyze the abnormal fields and associated detection device information corresponding to the abnormal data of the device, locate the root cause of the abnormality and output the device adjustment command;

[0125] S7: Valid data output and full lifecycle traceability. Receive the valid data frame, compile the valid data frame into a quality report, and construct a digital twin by associating the quality report, downstream processing data, and corrosion resistance data. The digital twin is used to feed back and update the dynamic judgment threshold.

[0126] In this embodiment, by constructing a parameter verification and data processing system specifically for zinc, aluminum, and magnesium, false scratches and invalid data are accurately filtered out, avoiding the problem of poor adaptability of general solutions, thus improving the accuracy of scratch data processing. Through bidirectional feedback between data processing and front-end devices, the quality of raw data collection is optimized from the source, reducing the generation of subsequent invalid data and ensuring the stability of production data. By linking the entire chain of data with a digital twin and feeding back into the processing rules, a closed-loop optimization of data from processing to application is achieved, continuously improving the system's long-term processing capabilities and meeting the needs of precise quality control in industrial production.

[0127] Based on the above embodiments, such as Figure 3 As shown, the present invention also provides an electronic device, the electronic device comprising:

[0128] The processor 22 includes at least one processor 22, at least one memory 21, a communication interface 23, and a communication bus 24, wherein the processor 22 is communicatively connected to the memory 21.

[0129] In this embodiment, the memory 21 can be implemented in any suitable manner, for example, the memory 21 can be a read-only memory, a hard disk drive, a solid-state drive, or a USB flash drive, etc.; the memory 21 is used to store at least one executable instruction executed by the processor;

[0130] In this embodiment, the processor 22 can be implemented in any suitable manner. For example, the processor 22 can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) that can be executed by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc.; the processor is used to execute the executable instructions to implement the zinc-aluminum-magnesium steel strip surface scratch data processing method as described above.

[0131] Based on the above embodiments, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for processing surface scratch data of zinc-aluminum-magnesium steel strip.

[0132] Those skilled in the art will recognize that the modules and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0133] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, equipment, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0134] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or equipment, and may be electrical, mechanical, or other forms.

[0135] The modules described as separate components may or may not be physically separate. The components shown as modules 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.

[0136] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0137] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion 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 described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program instructions, such as USB flash drives, portable hard drives, read-only storage servers, random access storage servers, magnetic disks, or optical disks.

[0138] Furthermore, it should be noted that the combination of the various technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0139] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

[0140] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A data processing system for surface scratches on zinc-aluminum-magnesium steel strip, characterized in that, include: The material parameter input module is used to receive and verify the basic parameters of zinc-aluminum-magnesium steel strip and output the parameter package that has passed the verification. The raw scratch data input module is used to receive the raw scratch data block output by the front-end detection device, perform integrity verification on the raw scratch data block, and output the raw data block that has passed the verification. A dedicated feature database for zinc, aluminum, and magnesium is used to store dynamic feature parameters associated with the properties of zinc, aluminum, and magnesium steel strips. After receiving the parameter package, the database retrieves and outputs the corresponding feature judgment criteria. The multi-source parameter association learning module is used to receive production process parameters and historical training data from the basic parameters and generate dynamic judgment thresholds; The dual dynamic verification module receives the original data block that has passed verification, the feature judgment benchmark and the dynamic judgment threshold respectively, performs grayscale verification and morphological verification on the original data block that has passed verification, and outputs a data frame with verification result mark through key judgment logic. The invalid data filtering module receives the data frame with the verification result mark, classifies the data frame with the verification result mark, and outputs valid data frames and device abnormal data. The bidirectional feedback correction module receives the abnormal data from the equipment, analyzes the root cause of the abnormal data, and outputs equipment adjustment instructions. The effective data output and digital twin module receives the effective data frame, outputs a quality report, and constructs a digital twin. The digital twin is used to associate the effective data frame, downstream processing data, and corrosion resistance data, and feeds back into the multi-source parameter association learning module. The zinc-aluminum-magnesium specific feature database includes: A three-dimensional index storage unit is used to store a tree-like index structure with coating thickness as the first-level index, magnesium content as the second-level index, and storage time as the third-level index. The storage time is divided into intervals of 0-24h, 24-48h, 48-72h, and greater than 72h. The oxidation model unit has a built-in oxide layer thickness calculation model, and its calculation formula is as follows: ,in, This indicates the thickness of the oxide layer, in μm, with a value ranging from 0 to 2.0 μm. The rate constant is expressed in units of 1000 ppm. The correlation between the magnesium content and the magnesium content is positive, specifically: when the magnesium content is 1.5%, =0.02; when the magnesium content is 2.0%, =0.025; when the magnesium content is 2.5%, =0.03; when the magnesium content is 3.0%, =0.04; This represents the storage time, in hours, and ranges from 0 to 168 hours. The feature benchmark output unit is used to receive the parameter package, calculate the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieve the three-dimensional index storage unit and call the oxidation model unit to correct the grayscale range, and output the feature judgment benchmark containing the grayscale range, grayscale compensation coefficient and depth-length rule to the dual dynamic verification module. The dual dynamic verification module includes: A data integration unit is used to integrate the verified original data block, the feature determination benchmark, and the dynamic determination threshold. The grayscale verification unit is used to calculate the storage time based on the collection timestamp in the verified original data block, call the oxidation model unit to calculate the oxide layer thickness to correct the grayscale range, and combine the light intensity to apply the grayscale compensation coefficient to determine whether the grayscale value in the verified original data block is within the corrected grayscale range. The morphological verification unit is used to extract the scratch depth and scratch length from the original data block that has passed the verification, calculate the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and determine whether the ratio is not less than 1. The key judgment unit is used to execute the key judgment logic to determine whether the original data block that has passed the verification has passed the verification of the grayscale verification unit and the morphological verification unit at the same time. If it passes at the same time, it is marked as data to be confirmed as valid data. If it does not pass at the same time, it is marked as invalid data to be classified. The unit outputs the data frame with the verification result mark containing the data to be confirmed as valid data and the data to be classified as invalid data.

2. The zinc-aluminum-magnesium steel strip surface scratch data processing system according to claim 1, characterized in that, The material parameter input module includes: The parameter receiving unit is used to receive the basic parameters of the zinc-aluminum-magnesium steel strip, including batch number, coating thickness, magnesium content, rolling completion time and production equipment number. The parameter verification unit is used to verify the format, value range and validity of the basic parameters. If the verification fails, an exception code is output and a preset processing action is executed. The preset processing action includes refusing to receive or triggering manual confirmation. The data interface unit uses a preset industrial Ethernet protocol to interface with the production execution system. The transmission rate of the preset industrial Ethernet protocol is not less than 100Mbps, and the parameter transmission delay is not more than 100ms. The parameter package output unit is used to encapsulate the verified basic parameters into the parameter package and send it synchronously to the original scratch data input module and the zinc-aluminum-magnesium exclusive feature database.

3. The zinc-aluminum-magnesium steel strip surface scratch data processing system according to claim 2, characterized in that, The parameter verification unit includes: The format verification subunit is used to verify whether the format of the batch number is a preset identifier + 6-digit date + 3-digit serial number, wherein the preset identifier is "ZM" and the total length of the batch number is fixed at 12 digits; The value verification subunit is used to verify whether the value range of the coating thickness is 5.0-15.0μm, whether the value range of the magnesium content is 1.5%~3.0%, whether the rolling completion time is not a future time, and whether the production equipment number is in the preset equipment list. An exception handling subunit is used to output a corresponding exception code when the basic parameters do not meet the verification requirements of the format verification subunit or the value verification subunit. The exception codes include P001 for batch number format exception, P002 for coating thickness exception, P003 for magnesium content exception, P004 for rolling completion time exception, and P005 for production equipment number exception.

4. The zinc-aluminum-magnesium steel strip surface scratch data processing system according to claim 1, characterized in that, The original scratch data input module includes: The data block receiving unit is used to receive the original scratch data block sent by the front-end detection device. The original scratch data block is encapsulated in JSON format. Every 100 frames of data constitute one original scratch data block. The header of the original scratch data block contains a batch number, a data block number, and a verification field for the number of frames. The integrity verification unit is used to verify whether the actual number of frames in the original scratch data block is consistent with the number of frames in the header, whether the detection area coordinates in the original scratch data block are not repeated, and whether the scratch depth and scratch length in the original scratch data block are non-negative values. The data block processing unit is used to output the original data block that has passed the integrity verification unit to the dual dynamic verification module when the original scratch data block passes the verification; when the original scratch data block fails the verification, it triggers a retransmission instruction, the maximum number of times the retransmission instruction is triggered is 3. If the verification still fails after 3 retransmissions, the original scratch data block is marked as an abnormal data block and sent to the invalid data filtering module.

5. The zinc-aluminum-magnesium steel strip surface scratch data processing system according to claim 1, characterized in that, The effective data output and digital twin module includes: The quality report generation unit is used to compile the valid data frames into a batch summary table and a list of valid scratch coordinates. The batch summary table includes the batch number, the total number of valid scratches, the number of scratches in different depth ranges, the maximum scratch size, and the data pass rate. The list of valid scratch coordinates includes the detection area coordinates, scratch depth, scratch length, and collection timestamp. A digital twin construction unit is used to associate the batch summary table, downstream processing data, and corrosion resistance data to construct the digital twin. The downstream processing data includes scratch extension records in the stamping process, and the corrosion resistance data includes salt spray test results. The data feedback unit is used to send the data in the digital twin to the multi-source parameter association learning module to trigger incremental training of the multi-source parameter association learning module and update the dynamic judgment threshold.

6. A method for processing surface scratch data of zinc-aluminum-magnesium steel strip, characterized in that, Includes the following steps: S1: Material parameters and raw data are input synchronously; receive and verify the basic parameters of zinc-aluminum-magnesium steel strip; output the parameter package that has passed verification; receive the raw scratch data block output by the front-end detection equipment; after verifying the integrity of the raw scratch data block, output the raw data block that has passed verification. S2: Dedicated feature parameter retrieval, receiving the parameter package, calculating the storage time based on the coating thickness, magnesium content and rolling completion time in the parameter package, retrieving preset dynamic feature parameters and correcting the grayscale range, and outputting the corresponding feature judgment benchmark; S3: Multi-source parameter association rule generation: Receive production process parameters and historical training data from the basic parameters, train the production process parameters and historical training data using a preset algorithm, and generate a dynamic judgment threshold. S4: Dual dynamic verification and key judgment, integrating the verified original data block, the feature judgment benchmark, and the dynamic judgment threshold, sequentially performing grayscale verification and morphological verification on the verified original data block, and outputting a data frame marked with the verification result through key judgment logic; wherein, the grayscale verification includes calculating the storage time based on the acquisition timestamp in the verified original data block, correcting the grayscale range by calling the oxide layer thickness calculation formula, applying a grayscale compensation coefficient in combination with the illumination intensity, and judging whether the grayscale value in the verified original data block is within the corrected grayscale range; the morphological verification includes extracting the scratch depth and scratch length in the verified original data block, calculating the ratio of the scratch length to the minimum length in the dynamic judgment threshold, and judging whether the ratio is not less than 1; S5: Invalid data filtering and classification: Receive the data frame with verification result mark, classify the data frame with verification result mark according to the preset classification rule, and output valid data frame and device abnormal data; S6: Two-way feedback correction: Receive the abnormal data of the device, analyze the abnormal fields and associated detection device information corresponding to the abnormal data of the device, locate the root cause of the abnormality and output the device adjustment command; S7: Valid data output and full lifecycle traceability. Receive the valid data frame, compile the valid data frame into a quality report, and construct a digital twin by associating the quality report, downstream processing data, and corrosion resistance data. The digital twin is used to feed back and update the dynamic judgment threshold.

7. An electronic device, characterized in that, The electronic device includes: The processor and the memory are communicatively connected. The memory is used to store at least one executable instruction executed by the processor, the processor being used to execute the executable instruction to implement the zinc-aluminum-magnesium steel strip surface scratch data processing method as described in claim 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for processing surface scratch data of zinc-aluminum-magnesium steel strip as described in claim 6.

Images