A rolling mill steel sleeve with an electronic identification chip and its data information management method

Abstract

This invention discloses a rolling mill steel sleeve with an electronic identification chip and its data information management method. Utilizing the continuity of time-series data and pattern matching technology, it transforms local single-point information into global state inference based on historical patterns, thereby solving the problem of logical disconnect due to insufficient data representativeness. This achieves accurate characterization of the overall structural state of the sleeve, supporting precise risk level assessment and scientific decision-making. The electronic identification chip reads the sensor data of the steel sleeve and integrates it into a unified time-series data record. By comparing real-time process parameters with a deformation mode feature library, it achieves comprehensive monitoring of rolling mill process parameters and efficient differentiation of deformation modes. A clustering analysis method based on parameter thresholds is used to establish a deformation mode feature library containing five typical patterns, enabling the summarization and classification of complex historical data, significantly improving the scientific nature of the feature library and the differentiation of similar deformation modes.

Description

Technical Field

[0001] This invention relates to the field of industrial production control technology, and more particularly to a rolling mill steel sleeve with an electronic identification chip and a method for managing its data information. Background Technology

[0002] In modern cold-rolled silicon steel strip production processes, the steel sleeve, as the rigid core in key processes such as coiling, annealing, and finishing, plays a crucial role in ensuring the strip's shape quality and preventing collapse, deformation, and surface damage due to its structural stability and traceability. With the metallurgical industry's deepening evolution towards automation and intelligence, the full lifecycle information management of this key component, the steel sleeve, has become a core technological aspect for improving production efficiency and process control precision.

[0003] In early technological practices, steel sleeves were typically made of high-strength alloy steel forgings, with precision machining ensuring their mechanical properties. Identification was achieved through physical methods such as engraving serial numbers on the sleeve ends. Operators had to manually record and verify the sleeve's weight, deformation, and associated steel coil information at close range. While this method met basic production management needs to some extent, its inherent drawbacks became increasingly apparent. These included low information recording efficiency, susceptibility to errors, and the significant safety risks posed by operators working under complex conditions of high tension and high temperature.

[0004] To overcome the aforementioned limitations, those skilled in the art have introduced electronic identification technology. A typical solution involves embedding an industrial-grade electronic chip in a pre-set groove on the end face of the steel sleeve. This chip stores the sleeve's unique identifier, key physical parameters, and usage history. Correspondingly, handheld or fixed reading devices can be used to remotely and non-contactly collect sleeve information and synchronize the data to the central management system in real time. This technological solution represents a significant breakthrough in the information management of steel sleeves. It successfully transforms manual operation into an automated data flow, greatly improving the efficiency and accuracy of information exchange, effectively avoiding safety hazards associated with close-range operations, and providing a preliminary data foundation for achieving full lifecycle tracking and intelligent scheduling of steel sleeves.

[0005] However, a deep-seated contradiction in the underlying principle of integrating single-point electronic chips onto steel sleeves has gradually become apparent. In its data acquisition and management logic, this approach implicitly simplifies the steel sleeve, a cylindrical structure bearing complex loads, into a rigid, uniform body, assuming that deformation data acquired at a single point can equivalently represent the structural state of the entire sleeve. This simplification is acceptable because it is easy to implement and was feasible in early scenarios where high precision was not required. However, the physical reality is that during high-speed rolling, the steel sleeve is subjected to the triple coupling of radial pressure, axial tension, and temperature gradient from the strip. Its deformation pattern is far from uniform, exhibiting a highly dynamic and complex non-uniform spatial distribution. While the deformation data measured by the chip installed at the end of the sleeve is accurate at its measurement point, it deviates significantly and unpredictably from the actual state of the highest-risk area within the sleeve. This deviation is completely masked by existing data management systems. The management system receives seemingly normal end deformation data within safety thresholds but is unaware that plastic deformation far exceeding safety limits may have occurred in the middle of the sleeve.

[0006] Therefore, while existing technologies have solved the problem of automated data acquisition for steel sleeves, they have failed to address the fundamental challenge of how to accurately and effectively represent the global structural state of the sleeve under complex working conditions using collected local information from a single point. This logical disconnect in data representativeness prevents truly accurate prediction and scientific decision-making. Consequently, developing a new technical solution to overcome the limitations of single-point measurement and establish a dynamic mapping relationship between local measurement data and the overall deformation field of the sleeve, thereby providing comprehensive and accurate status assessment data for the upper-level management system, has become a key technical bottleneck that urgently needs to be addressed by those skilled in the art. Summary of the Invention

[0007] This invention overcomes the shortcomings of the prior art and provides a rolling mill steel sleeve with an electronic identification chip and its data information management method, aiming to solve the problem of how to truly reflect the deformation of the rolling mill steel sleeve in the prior art.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a data information management method for rolling mill steel sleeves with electronic identification chips, comprising the following steps:

[0009] S1: Read the sensor data on the steel sleeve, obtain the rolling mill process parameters, and integrate them into a unified time-series data record;

[0010] S2: Based on the time-series data records, the deformation modes are distinguished, and a deformation mode feature library is established; wherein, the deformation mode feature library includes 5 deformation modes: saddle-shaped deformation mode, conical deformation mode, spiral deformation mode, local concave deformation mode, and uniform deformation mode.

[0011] S3: Extract real-time process parameters from time-series data records, compare them with the deformation mode feature library, and identify deformation modes that conform to the current working conditions.

[0012] S4: Assess the risk level based on the deformation mode under the current working conditions.

[0013] In a preferred embodiment of the present invention, the sensing data on the steel sleeve is specifically the physical response of the steel sleeve body during the rolling process;

[0014] The rolling mill process parameters specifically refer to the control and process parameters during the operation of the rolling mill.

[0015] In a preferred embodiment of the present invention,

[0016] The sensing data in S1 includes the unique identifier of the steel sleeve and the radial deformation at the end. The radial deformation at the end includes the first end deformation and the second end deformation. The difference between the radial deformation at the end is the difference between the first end deformation and the second end deformation.

[0017] The mill process parameters in S1 include: real-time rolling speed, strip tension setpoint, strip target thickness, main rolling force, and cooling water flow rate;

[0018] The integration method is an interpolation algorithm based on a sliding time window.

[0019] In a preferred embodiment of the present invention, the establishment of the deformation pattern feature library in step S2 adopts a clustering analysis inductive model, specifically as follows:

[0020] Establish threshold conditions for process parameters of several of the aforementioned deformation modes;

[0021] Time series data records are classified based on process parameter threshold triggering conditions;

[0022] Centroids are calculated for time-series data records in each category, and clustering is performed based on the centroids.

[0023] In a preferred embodiment of the present invention, the threshold values ​​in the deformation mode feature library are as follows:

[0024] The threshold conditions for the saddle-shaped deformation mode are: the main rolling force is greater than 18000kN and the strip thickness is less than 1.5mm;

[0025] The threshold conditions for the conical deformation mode are: the strip tension setting value is greater than 5kN and the difference in radial deformation at the ends is greater than 0.1mm;

[0026] The threshold conditions for the spiral deformation mode are: real-time rolling speed greater than 0.5 m / s and end radial deformation change rate greater than 0.05 mm / s;

[0027] The threshold condition for the local indentation deformation mode is: cooling water flow rate greater than 50. Furthermore, the peak value of the radial deformation at the end is greater than 0.3 mm;

[0028] Uniform deformation mode threshold condition: triggered when the real-time parameters do not meet the above threshold.

[0029] In a preferred embodiment of the present invention, the comparison process with the deformation pattern feature library in S3 includes threshold screening and weighted Euclidean distance precise matching, the specific steps of which are as follows:

[0030] Compare real-time process parameters with parameter thresholds;

[0031] Determine if it matches the parameter threshold;

[0032] Based on the judgment result, determine whether to perform weighted Euclidean distance exact matching.

[0033] In a preferred embodiment of the present invention, the criterion for determining the weighted Euclidean distance exact matching is:

[0034] If the parameter threshold matches any one or more of the saddle-shaped deformation mode, cone-shaped deformation mode, spiral deformation mode, and local concave deformation mode, then a weighted Euclidean distance is used for precise matching.

[0035] Otherwise, it is determined to be a uniform deformation mode.

[0036] In a preferred embodiment of the present invention, the weighted Euclidean distance precise matching specifically involves quantizing and comparing the real-time process parameter vector with the parameterized trigger condition set of all modes.

[0037] In a preferred embodiment of the present invention, the risk levels in S4 include high risk, medium risk, and low risk, and the deformation modes of the current working condition correspond to the respective risk levels; wherein,

[0038] The high-risk pattern is the saddle-shaped deformation mode;

[0039] Medium-risk conditions include conical deformation mode, spiral deformation mode, and localized indentation deformation;

[0040] Low risk is the uniform deformation mode.

[0041] To achieve the above objectives, the second technical solution adopted by the present invention is as follows: a rolling mill steel sleeve with an electronic identification chip, comprising a sleeve body, wherein blind mounting holes are provided on both end faces of the sleeve body, and a smart sensor chip module is fixedly sealed within the blind mounting holes. The smart sensor chip module is an integrated microelectronic system, which internally encapsulates:

[0042] The identity recognition unit is used to permanently store the unique identity and attribute information of the steel sleeve;

[0043] The deformation sensing unit is used to measure the local deformation and vibration state of the cylinder end face under working load in real time.

[0044] The data processing and storage unit is used to process, calculate and store the signals collected by the deformation sensing unit;

[0045] A wireless transceiver unit is used for wireless data exchange with external reading and writing devices; and

[0046] An energy harvesting unit is used to harvest the vibration energy of the steel sleeve during operation and convert it into electrical energy;

[0047] The identity recognition unit, deformation sensing unit, data processing and storage unit, and wireless transceiver unit are electrically connected in sequence, and the energy harvesting unit provides working power to the other units in the intelligent sensing chip module.

[0048] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0049] This invention provides a data information management method for rolling mill steel sleeves with electronic identification chips. By utilizing the continuity of time-series data and pattern matching technology, local single-point information is transformed into global state inference based on historical patterns. This solves the problem of logical disconnect due to insufficient data representativeness, and achieves accurate characterization of the global structural state of the sleeve. This supports accurate risk level assessment and scientific decision-making. Compared with existing technologies, this method reads the sensor data of the steel sleeve through electronic identification chips and integrates it into a unified time-series data record. By comparing and identifying real-time process parameters with a deformation mode feature library, it achieves comprehensive monitoring of rolling mill process parameters and efficient differentiation of deformation modes.

[0050] In this invention, a clustering analysis method based on parameter thresholds is used to establish a feature library of deformation patterns containing five typical patterns. This enables the summarization and classification of complex historical data. By dividing the multidimensional process parameter dataset through cluster centroids, expert experience or historical patterns are transformed into a parameterized and quantifiable set of pattern triggering conditions. Compared with existing technologies, this transforms the identification of deformation patterns from qualitative judgment to quantitative classification based on data distribution, significantly improving the scientific nature of the feature library and its ability to characterize complex nonlinear relationships.

[0051] In this invention, weighted Euclidean distance calculation is used to assign weights to different process parameters that reflect their degree of influence. This allows for a more accurate measurement of the similarity between actual working conditions and various preset modes in quantitative comparisons, quickly identifying the most likely deformation mode. A feature library containing multiple typical deformation modes is preset, and a weighted Euclidean distance algorithm is used for identification and matching, improving the accuracy and efficiency of mode matching. Compared with existing technologies, this invention quantitatively compares the real-time acquired process parameter vectors with the standard parameterized triggering conditions of various modes in the feature library, enabling the system to predict the unique risk evolution path of the identified specific modes in advance.

[0052] In this invention, the identified specific deformation patterns are directly associated with high-risk, medium-risk, and low-risk levels, enabling qualitative risk assessment of the current working condition. Different deformation patterns correspond to different stress distribution anomalies and failure mechanisms, which are directly mapped to risk levels through pattern recognition. Compared with existing technologies, this transforms the complex structural mechanics state assessment problem into a traceable analogical reasoning process based on historical data and a pattern library, thereby achieving a rapid, accurate, and interpretable assessment of the overall structural state risk of the steel sleeve.

[0053] In this invention, the deformation types of the steel sleeve are dynamically combined with the radial deformation at the end and the rolling mill process parameters. The end deformation provides real feedback on local deformation, while the process parameters reflect the global load distribution, reconstructing the deformation state of the entire sleeve. The multi-source data coupling mechanism overcomes the defect that single-point measurement cannot capture spatial deformation, enabling the system to accurately identify specific patterns and trigger early warnings based on composite risk rules. Ultimately, it achieves closed-loop optimization from data acquisition to risk intervention. Compared with existing technologies, it improves the accuracy and real-time performance of deformation pattern recognition, avoids misjudgment from a single data source, integrates the radial deformation at the end and the rolling mill process parameters through timestamp alignment, and uses a weighted Euclidean distance algorithm to compare with a preset feature library in real time, improving production safety and the scientific nature of decision-making. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the method steps of a preferred embodiment of the present invention;

[0056] Figure 2 This is a perspective view of a preferred embodiment of the present invention;

[0057] Figure 3 This is a perspective view of a preferred embodiment of the present invention, specifically Embodiment 2.

[0058] In the attached diagram: 1. Sleeve body; 2. Chip pressure plate; 3. Electronic chip. Detailed Implementation

[0059] 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.

[0060] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0061] Application Overview:

[0062] While existing technologies using rolling mill sleeves equipped with electronic identification chips have achieved automated information acquisition, the acquired information is often single-point, localized, and static. This method cannot effectively capture the overall structural response and evolution of the sleeve under the complex coupling effects of rolling forces, thermal stresses, and other fields, resulting in a serious logical disconnect between the acquired data and the actual global structural state of the sleeve. This fundamental lack of data representativeness renders lifespan prediction and health management based on this data lacking in scientific basis, hindering accurate prediction and scientific decision-making.

[0063] The reason why the true global structural state of the steel sleeve cannot be monitored is that the sensing data in the middle of the steel sleeve is difficult to monitor, and the data detected from the ends of the steel sleeve cannot fully reflect the sensing situation in the middle of the steel sleeve. As a result, the middle of the steel sleeve is the direct contact area of ​​the strip winding, which is subjected to extremely high radial pressure and strong friction and wear. If a sensor is installed in the middle, it is necessary to make holes or embed modules, which will significantly weaken the structural strength of the sleeve and easily cause stress concentration or even fracture.

[0064] Unexpectedly, it was discovered that the main types of deformation in steel sleeves are saddle-shaped deformation, conical deformation, spiral deformation, and local dents. During the operation of these deformed steel sleeves, there is a correlation between the radial deformation at the end and the rolling mill process parameters. Therefore, by combining the types of deformation of the steel sleeve with the radial deformation at the end and the rolling mill process parameters, it is possible to obtain accurate feedback on the overall deformation degree of the steel sleeve without measuring the sensor data in the middle of the steel sleeve.

[0065] This application simultaneously reads steel sleeve sensor data and rolling mill process parameters, and integrates them into unified time-series data through sliding window interpolation. Based on historical data, it uses cluster analysis to summarize five typical deformation modes, including saddle-shaped and conical, and constructs a parameterized feature library. For real-time data, it performs rapid initial screening of parameter thresholds and performs weighted Euclidean distance precise matching on suspected non-uniform deformation conditions to identify the most matching deformation mode. Based on the identified deformation mode, it determines its corresponding risk level.

[0066] Overcoming the limitations of single-point information monitoring, this approach places instantaneous, localized monitoring data within a continuous process timeline and matches it with a feature pattern library representing different global deformation mechanisms. This essentially enables data-driven deduction of the overall mechanical state of the cylinder. This allows decision-makers to anticipate the development trends of specific failure modes, such as saddle-shaped deformation, thereby upgrading maintenance strategies from passive response to proactive risk prevention. This significantly improves equipment safety and operational efficiency and provides core decision support for predictive maintenance.

[0067] This invention provides a data information management method for rolling mill steel sleeves with electronic identification chips, comprising the following steps:

[0068] S1: Reads sensor data from the steel sleeve to obtain rolling mill process parameters and integrates them into a unified time-series data record; uses a reading and writing device to periodically or in real-time read the identification information stored in the electronic chip on the steel sleeve, and simultaneously collects data from various sensors installed on the rolling mill or sleeve body. By combining static identification information with dynamic sensor data and uniformly timestamping it, a complete and continuous time-series data record of the sleeve's working process is formed, ensuring that the data for each steel sleeve is traceable throughout its entire process from commissioning, operation to maintenance.

[0069] S2: Based on time-series data records, deformation patterns are differentiated, and a deformation pattern feature library is established. Data analysis and machine learning algorithms are used to deeply explore the multi-dimensional data characteristics exhibited by steel sleeves under the rolling process. These features are patterned and labeled to establish a deformation pattern feature library containing multiple typical working conditions. Complex, unstructured raw data is refined into standardized patterns with clear engineering significance, providing a comparison benchmark and knowledge base for subsequent real-time pattern recognition.

[0070] S3: Extract real-time process parameters from the time-series data records and compare them with the deformation mode feature library to identify the deformation mode that matches the current operating conditions. Extract the real-time process parameter sequence of the steel sleeve within the current rolling cycle online and use a pattern recognition algorithm to quickly and accurately compare and match it with the deformation mode feature library established in S2. This allows for immediate determination of the known deformation mode currently in which the steel sleeve is operating. This not only helps in understanding the current operating status but also provides early warnings of impending wear or deformation trends, transforming passive response into proactive perception.

[0071] S4: Assess the risk level based on the deformation mode under the current working conditions. After identifying the current deformation mode, the system will assess the current health status of the sleeve and classify the risk level.

[0072] By leveraging the continuity of time-series data and pattern matching technology, local single-point information is transformed into global state inference based on historical patterns. This solves the problem of logical disconnect due to insufficient data representativeness, enabling accurate characterization of the global structural state of the cylinder. This supports precise risk level assessment and scientific decision-making. The sensor data of the steel sleeve is read by an electronic identification chip and integrated into a unified time-series data record. By comparing and identifying real-time process parameters with the deformation mode feature library, comprehensive monitoring of rolling mill process parameters and efficient differentiation of deformation modes are achieved.

[0073] The steps will be described in detail below:

[0074] In step S1,

[0075] The sensing data on the steel sleeve specifically refers to the physical response of the steel sleeve body during the rolling process, while the rolling mill process parameters specifically refer to the control and process parameters during the operation of the rolling mill.

[0076] The sensing data includes a unique identifier for the steel sleeve and the radial deformation at the end, which includes a first end deformation and a second end deformation. The difference between the two end radial deformations is the difference between the first end deformation and the second end deformation.

[0077] The unique identifier of a steel sleeve is a unique sequence of numbers or characters used to index all historical data of the steel sleeve in the database; the radial deformation at the end represents the change in the diameter direction of the steel sleeve end under the action of rolling force. It is a key direct indicator for assessing its mechanical load and health status, and serves as objective evidence to verify whether all process parameters ultimately lead to deformation. Without this parameter, the condition assessment loses its factual basis.

[0078] The rolling mill process parameters include: real-time rolling speed, strip tension setpoint, strip target thickness, main rolling force, and cooling water flow rate; the alignment of the two types of data is based on the interpolation algorithm of the sliding time window.

[0079] Real-time rolling speed reflects the speed of the rolling process. Speed ​​changes will cause equipment vibration and dynamic load changes, creating alternating stress on the sleeve and aggravating fatigue. The rolling process generates a large amount of deformation heat and frictional heat. The faster the rolling speed, the more heat is generated per unit time, providing an energy basis for thermal deformation.

[0080] The strip tension setting is a crucial process parameter for ensuring smooth and unbiased strip rolling. During rolling, the strip needs to maintain a certain tension at the inlet and outlet. This tension translates into a horizontal pulling force on the rolls, thus affecting the stress state of the rolls and the internal steel sleeves. Uneven or excessive tension introduces asymmetrical loads, which are significant factors inducing conical or helical deformation.

[0081] The target thickness of the strip is one of the main quality objectives pursued in strip rolling, which is a plastic deformation process. To change its thickness, the rolls must apply sufficient force to overcome the yield strength of the strip material. The smaller the target thickness, the greater the reduction from the inlet thickness to the outlet thickness, and the greater the deformation work required. Therefore, the main rolling force must be increased proportionally. Extremely high main rolling forces directly generate enormous radial pressure on the support rolls and the steel sleeves inside, inducing deformation.

[0082] The main rolling force is the total pressure applied by the rolling mill to the strip and the sleeve. It is the primary source of load, and the magnitude of the main rolling force directly determines the level of extrusion stress on the sleeve. Excessive main rolling force is a major cause of uniform wear, plastic deformation, and even localized crushing.

[0083] Cooling water flow rate is used to control the temperature of the rolls and strip, affecting rolling quality and roll wear. Insufficient or uneven flow rate can lead to excessively high local temperatures on the rolls, generating significant thermal stress. This thermal stress, combined with mechanical stress, can easily cause deformation.

[0084] During the rolling process, the steel sleeve is subjected to the triple coupling of radial pressure, axial tension, and temperature gradient from the strip, resulting in a highly dynamic and complex non-uniform spatial distribution of deformation. Post-processing management relies on single-point measurement data, assuming the sleeve is rigid and uniform. However, the condition of the highest-risk area—the center of the sleeve—is completely masked. For example, during saddle-shaped deformation, the end data may appear normal, but the center may have already undergone plastic deformation far exceeding safety limits, making it impossible to trace the true risk in post-processing analysis.

[0085] By integrating real-time sensor data with rolling process parameters, multi-source data alignment and dynamic matching of deformation modes are achieved. This real-time capability enables the system to infer the overall state with high precision based on local measurements, preventing the escalation of accidents.

[0086] During mill operation, the radial deformation at the end and mill process parameters are collected synchronously and integrated in real time for analysis. This avoids the blind spots of single-point static measurements and improves the authenticity and timeliness of the data. The radial deformation at the end and mill process parameters are merged into unified time-series data through timestamp alignment, and then dynamically matched with a preset deformation mode feature library using a weighted Euclidean distance algorithm.

[0087] End data is given spatial meaning, and key coefficients are solved by instantiating the mathematical model of deformation field. This allows for high-precision reconstruction of the overall deformation state of the cylinder from local measurements, overcoming the limitation that single-point data cannot capture complex deformation patterns and enabling early risk warning and decision optimization.

[0088] The reading frequency of steel sleeve sensor data and the sampling frequency of rolling mill process parameters are usually out of sync. Their timestamps cannot be precisely matched, and direct merging will lead to data misalignment or significant data loss. Therefore, an algorithm is needed to solve the problem of timestamp inconsistency.

[0089] Using the timestamps within the data packets as a reference, an interpolation algorithm based on a sliding time window is used to align asynchronous data from the steel sleeve and the rolling mill, ultimately resulting in a structured wide table data record.

[0090] The specific steps are as follows: Set a reasonable time tolerance window w, and for each steel sleeve data point, calculate the starting point of its time window. and the end point This window defines the range of relevant rolling mill data we are looking for.

[0091] Find the timestamp in the rolling mill process parameter data. Falling All data points within the interval. For each process parameter p that needs alignment, find the distance within the window. The two most recent rolling mill data points:

[0092] point : ,in It is less than The latest timestamp.

[0093] Point β: ,in It is greater than The earliest timestamp.

[0094] Using linear interpolation, the estimation is performed at... Process parameter values ​​at time The calculation formula is: .

[0095] The calculated All process parameter values ​​at any given time are combined with the unique identification code of the steel sleeve and the radial deformation at the end to form a final, timestamp-aligned data record. The feature vector of the data point is [real-time rolling speed, strip tension setpoint, strip target thickness, main rolling force, cooling water flow rate, and end deformation].

[0096] By synchronously collecting and integrating the physical response of the steel sleeve body with multiple process control parameters of the rolling mill, multi-dimensional time-series data covering equipment status and operating conditions were obtained. Using the sliding time window interpolation algorithm, the two types of data from different sources and with different frequencies were forced to be precisely aligned on the time axis, eliminating the characterization distortion caused by data asynchrony. This allowed the originally isolated sleeve deformation data to be placed in a complete process context, laying a data foundation for accurately establishing the causal relationship between deformation and operating conditions.

[0097] In step S2,

[0098] The deformation mode feature library includes saddle-shaped deformation mode, conical deformation mode, spiral deformation mode, local concave deformation mode, and uniform deformation mode.

[0099] The deformation pattern feature library was established using a clustering analysis and inductive approach. Specifically, the process parameter datasets corresponding to the five deformation patterns were assigned based on parameter thresholds, and centroids were assigned for clustering.

[0100] In the saddle-shaped deformation mode, the deformation in the middle of the cylinder is greater than that at both ends; in the conical deformation mode, the deformation at one end of the cylinder is significantly greater than that at the other end; in the spiral deformation mode, the deformation of the cylinder is distributed spirally along the axial direction; in the local concave deformation mode, the local area is concave; in the uniform deformation mode, the cylinder still maintains a uniformly distributed cylindrical shape.

[0101] The thresholds in the deformation pattern feature library are as follows:

[0102] Threshold conditions for saddle-shaped deformation mode: main rolling force greater than 18000kN and strip thickness less than 1.5mm; the essence of saddle-shaped deformation is that the radial pressure in the middle region of the cylinder is significantly greater than that at both ends. When rolling thin strip steel, a higher rolling force is required to achieve the target thickness, causing stress to concentrate in the middle of the cylinder. When thin strip steel is combined with a high rolling force, plastic deformation is prone to occur in the middle of the cylinder, leading to the appearance of the saddle-shaped deformation mode.

[0103] Based on finite element simulation data, the simulation shows that when the rolling force exceeds 18000kN, the stress concentration factor in the middle of the cylinder increases significantly, resulting in a saddle-shaped deformation probability exceeding 90%. Historical data statistics show that the rolling force is higher than 17500kN in 95% of saddle-shaped events. Setting 18000kN provides a 5% safety margin to avoid false alarms caused by fluctuations in operating conditions.

[0104] Thin strip steel requires higher rolling forces, which exacerbates the deflection in the middle. Simulation analysis shows that when the thickness is less than 1.5 mm, the deformation morphology of the cylinder transitions from uniform to saddle-shaped, and this value is the statistical median of the material's deformation critical point.

[0105] Threshold conditions for conical deformation mode: The strip tension setting value is greater than 5kN and the difference in radial deformation at the ends is greater than 0.1mm; conical deformation is caused by uneven axial load. The tension gradient causes one end of the cylinder to bear greater tensile stress. Combined with the difference in deformation at the ends, the degree of asymmetry can be directly quantified, and the conical deformation mode can be directly displayed.

[0106] Historical data analysis shows that when the tension gradient exceeds 5kN, the probability of uneven axial load leading to conical deformation reaches 85%. This value is based on the data distribution of the rolling mill tension sensor to ensure the capture of typical deviation conditions; the 95th percentile of the difference in deformation at both ends during conical deformation is 0.08mm, and a tolerance of 0.1mm is set to provide a margin of error and avoid noise interference.

[0107] Threshold conditions for spiral deformation mode: real-time rolling speed greater than 0.5 m / s and end radial deformation change rate greater than 0.05 mm / s; spiral deformation involves dynamic torsion, and speed fluctuations during high-speed rolling can easily cause torsional vibration. The end deformation change rate can capture this dynamic characteristic, and the spiral deformation mode can be observed more clearly when combined.

[0108] Simulation analysis shows that when the speed is below 0.5 m / s, deformation is mainly static, while above this value, the dynamic torque effect increases, and the risk of spiral deformation increases. This value is 50% of the mill's rated speed, based on vibration frequency matching; the rate of change of end deformation is calculated using historical vibration data, and when the rate of change is >0.05 mm / s, torsional vibration characteristics are obvious.

[0109] Threshold condition for local indentation deformation mode: Cooling water flow rate greater than 50 Furthermore, the peak value of the radial deformation at the end is greater than 0.3 mm; local depressions are caused by concentrated loads or thermal stress. Uneven cooling leads to local thermal stress, and by combining the peak value of the end deformation to locate the depression area, the local depression deformation mode can be easily detected.

[0110] Historical data shows that the cooling flow deviation exceeds 50. At this time, the probability of local thermal stress causing indentation is relatively high. This value is the statistical upper limit of the flow range of the laminar flow cooling system to ensure the capture of abnormal operating conditions; based on the mechanical properties of the material, the critical value of local yield deformation of the steel sleeve material is about 0.25mm, and setting 0.3mm provides a safety margin to avoid plastic deformation failure.

[0111] Uniform deformation mode threshold condition: Triggered when none of the real-time parameters meet the above threshold. When all parameters are within the threshold, the load distribution is uniform and the deformation is within a safe range.

[0112] The process of assigning centroids for clustering involves classifying and organizing several time-series data records, and then calculating the centroids of the corresponding time-series data records in the modified mode. The centroid is the mean of the time-series data records in the cluster.

[0113] After completing the initial data partitioning based on thresholds, K-means clustering algorithm is used to calculate and optimize the centroid for each partitioned data subset. The vector average of all feature vectors in the historical data subset corresponding to each deformation pattern is calculated to obtain an initial centroid vector.

[0114] Starting from all initial centroid vectors, iterative clustering is performed on the entire historical time series database until the centroid positions of each cluster converge. The resulting five stable centroid vectors, together with the initially set parameter threshold rule set, constitute the deformation pattern feature library.

[0115] The centroid calculation process is as follows:

[0116] Process parameters and sensor data were collected from finite element simulation and historical measured data, including rolling speed, strip tension, thickness, main rolling force, cooling water flow rate and end radial deformation.

[0117] Select a set of process parameter data that only satisfies this cluster and use it as the initial centroid;

[0118] The new process parameter data that only satisfy this cluster will be selected according to the formula. The centroid is recalculated and iterated, where x is the data point and t is the number of iterations. It is clustering The number of data points in the data;

[0119] This continues until the centroids of all process parameter data that satisfy only this cluster are summarized from finite element simulation and historical measured data.

[0120] Specifically, process parameter data that satisfies only this cluster refers to process parameter data that meets the process parameter threshold of this cluster, and all parameters except the corresponding parameter do not meet the process parameter thresholds of other clusters. For example, in process parameter data that only satisfies the saddle-shaped deformation mode, the main rolling force is greater than 18000kN and the strip thickness is less than 1.5mm, and each of the following—real-time rolling speed, strip tension setpoint, cooling water flow rate, and end radial deformation—does not meet the process parameter thresholds of the rest of this cluster.

[0121] A feature library of deformation patterns containing five typical patterns was established by using a clustering analysis method based on parameter thresholds. This enabled the summarization and classification of complex historical data. By dividing the multidimensional process parameter dataset through cluster centroids, expert experience or historical patterns were transformed into a set of parameterized and quantifiable pattern triggering conditions. This transformed the identification of deformation patterns from qualitative judgment to quantitative classification based on data distribution, significantly improving the scientific nature of the feature library and its ability to characterize complex nonlinear relationships.

[0122] In step S3,

[0123] The comparison includes threshold screening and weighted Euclidean distance precise matching. Real-time process parameters are compared with parameter thresholds to determine whether they match the parameter thresholds. Based on the judgment results, it is determined whether to perform weighted Euclidean distance precise matching.

[0124] The criteria for determining weighted Euclidean distance precision matching are as follows: if the parameter thresholds match any of the deformation modes (saddle-shaped, conical, spiral, and local concave), weighted Euclidean distance precision matching is performed; otherwise, it is determined to be a uniform deformation mode.

[0125] Weighted Euclidean distance precise matching specifically involves quantifying and comparing the real-time process parameter vector with the parameterized trigger condition set of all modes. Before quantization, the parameters are processed using the Z-score standard, and unit conflicts are not considered during the comparison process. Weights are assigned based on parameter importance. The real-time rolling speed is in m / s, the strip tension setting is in kN, the target strip thickness is in mm, the main rolling force is in kN, and the cooling water flow rate is... The end deformation is in mm.

[0126] Let the real-time parameter vector be... , where n is the number of parameters, and in this application, the number of n is 7.

[0127] The deformable pattern feature library pre-stores the centroid of each pattern. Let the centroid of pattern m be... , where m can take the form of a saddle shape, a cone shape, etc. These vectors are derived based on finite element simulation and historical data.

[0128] Different process parameters have varying degrees of influence on deformation, therefore the algorithm assigns weights to each parameter. The weights are calculated based on the Pearson correlation coefficient between parameters and deformation patterns in historical data, using the following formula: ,in This is the correlation coefficient between parameter i and the mode correlation. In this application, the real-time rolling speed, strip tension setpoint, strip target thickness, main rolling force, cooling water flow rate, and end deformation are... The ranges are 0.70-0.80, 0.50-0.60, 0.30, -0.40, 0.25-0.35, 0.15-0.25, and 0.10-0.20, respectively.

[0129] Calculate the real-time parameter vector P and the feature vector of each mode. Weighted Euclidean distance The formula is: ,in, It is the weight of the i-th parameter. These are real-time parameter values. These are the eigenvalues ​​of pattern m. The unit is the reciprocal of the corresponding real-time parameter vector; for example, the unit of the weight of the strip tension setpoint is kN. -1 .

[0130] Choose to make distance The smallest mode is considered the variant mode of the current working condition, and the smaller the distance, the more suitable it is for the current working condition.

[0131] When the real-time parameter vector simultaneously satisfies two or more of the parameter thresholds corresponding to the saddle-shaped deformation mode, conical deformation mode, spiral deformation mode, and local concave deformation mode, the weighted Euclidean distance calculation and centroid iteration calculation are performed first. The specific steps are as follows:

[0132] The real-time vector is precisely matched with the corresponding deformation mode using a weighted Euclidean distance.

[0133] Determine the deformation mode that has the smallest weighted Euclidean distance to the real-time vector, and assign the real-time vector to this deformation mode.

[0134] The new deformation mode centroid is obtained by iterating over the real-time vector and the original deformation mode centroid.

[0135] By using weighted Euclidean distance calculation, different process parameters are assigned weights that reflect their degree of influence, thereby more accurately measuring the similarity between actual working conditions and various preset modes in quantitative comparison, and quickly locking in the most likely deformation mode. A feature library containing a variety of typical deformation modes is preset and a weighted Euclidean distance algorithm is used for identification and matching, which improves the accuracy and efficiency of mode matching. The real-time acquired process parameter vector is quantitatively compared with the standard parameterized trigger conditions of various modes in the feature library, which enables the system to predict the unique risk evolution path of the identified specific modes in advance.

[0136] By dynamically combining the deformation types of the steel sleeve with the radial deformation at the end and the rolling mill process parameters, the end deformation provides a true feedback on local deformation, while the process parameters reflect the global load distribution, thus reconstructing the deformation state of the entire sleeve. The multi-source data coupling mechanism overcomes the deficiency of single-point measurement in capturing spatial deformation, enabling the system to accurately identify specific patterns and trigger early warnings based on composite risk rules. Ultimately, it achieves closed-loop optimization from data acquisition to risk intervention, improving the accuracy and real-time performance of deformation pattern recognition and avoiding misjudgments from a single data source. By aligning and integrating the radial deformation at the end and the rolling mill process parameters through timestamps, and using a weighted Euclidean distance algorithm to compare with a preset feature library in real time, it improves production safety and the scientific nature of decision-making.

[0137] In step S4,

[0138] The risk levels are divided into high risk, medium risk, and low risk, and the deformation mode of the current working condition corresponds to the respective risk level.

[0139] In the risk level classification, high risk is saddle-shaped deformation mode, medium risk includes conical deformation mode, spiral deformation mode and local concave deformation, and low risk is uniform deformation mode.

[0140] When the risk level is high, the rolling process should be immediately suspended and maintenance carried out; when the risk level is medium, monitoring should be strengthened and process parameters should be optimized; when the risk level is low, no intervention is required.

[0141] The specific deformation patterns identified are directly associated with high-risk, medium-risk, and low-risk levels, enabling the qualitative assessment of the current working conditions. Different deformation patterns correspond to different stress distribution anomalies and failure mechanisms. By directly mapping the pattern recognition to the risk level, the complex problem of structural mechanics state assessment is essentially transformed into a traceable analogical reasoning process based on historical data and a pattern library. This enables a rapid, accurate, and interpretable assessment of the overall structural state risk of the steel sleeve.

[0142] A rolling mill steel sleeve with an electronic identification chip includes a sleeve body. Both end faces of the sleeve body have blind mounting holes. A smart sensor chip module is fixedly sealed within each blind mounting hole. The smart sensor chip module is an integrated microelectronic system, internally encapsulated with:

[0143] The identification unit is used to store the unique identity and attribute information of the steel sleeve; it is a passive radio frequency identification chip, specifically the UCODE8 chip from NXP.

[0144] The deformation sensing unit is used to measure the local deformation and vibration state of the cylinder end face under working load in real time; it consists of a piezoresistive strain gauge and a triaxial MEMS accelerometer.

[0145] The data processing and storage unit is used to process, calculate, and store the signals collected by the deformation sensing unit; it is a microcontroller.

[0146] The wireless transceiver unit, used for wireless data exchange with external reading and writing devices, is a radio frequency front-end circuit; and

[0147] The energy harvesting unit is used to collect the vibration energy of the steel sleeve during operation and convert it into electrical energy; it is a cantilever beam piezoelectric vibration energy harvester.

[0148] The identity recognition unit, deformation sensing unit, data processing and storage unit, and wireless transceiver unit are electrically connected in sequence, and the energy harvesting unit provides working power to the other units in the intelligent sensing chip module.

[0149] In this application, the sensor module is installed in a blind hole on the end face of the cylinder. This design avoids damaging the main load-bearing structure of the cylinder. The end position structure is relatively stable, and the space allows for the machining of blind holes, while the middle position cannot provide such installation conditions. By combining the end position sensing data with real-time operating condition data and establishing a deformation mode library, it is possible to ensure the real-world situation without affecting the rolling process.

[0150] In addition to the intelligent sensor chip module, external reading and writing devices and related processing electronic devices are also required.

[0151] An external reader / writer reads the sensor data from the steel sleeve chip module. The chip module itself is a passive design and relies on the reader / writer to activate and receive data.

[0152] Related processing electronic devices, including:

[0153] Processor; a processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities.

[0154] The memory stores computer program instructions, which, when executed by the processor, cause the processor to perform a data information management method for the rolling mill steel sleeve with an electronic identification chip.

[0155] The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory and / or cache memory. The non-volatile memory may include, for example, read-only memory, hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and a processor may execute the program instructions to implement the data information management method for rolling mill steel sleeves with electronic identification chips described in the various embodiments of this application above, and / or other desired functions.

[0156] A computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform a data information management method for a rolling mill steel sleeve with an electronic identification chip.

[0157] Computer-readable storage media can take the form of any combination of one or more readable media. A readable medium can be a readable signal medium or a readable storage medium. A readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0158] Example 1:

[0159] like Figure 2 As shown, a steel sleeve has a cylindrical body with a blind hole of a predetermined depth on both end faces of the cylindrical body. A smart sensor chip module is fixedly sealed inside the blind hole. Specifically, the sleeve body 1 and the electronic chip 3 are connected by bolts, and the electronic chip 3 has a chip pressure plate.

[0160] Example 2:

[0161] like Figure 3 As shown, a steel sleeve has a cylindrical body with a blind hole of a predetermined depth on both end faces of the cylindrical body. A smart sensor chip module is fixedly sealed inside the blind hole. Specifically, the sleeve body 1 and the electronic chip 3 are fixed together by using glue.

[0162] Based on the preferred embodiments of the present invention described above, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for managing data information of rolling mill steel sleeves with electronic identification chips, comprising the following steps: S1: Read the sensor data on the steel sleeve, obtain the rolling mill process parameters, and integrate them into a unified time-series data record; the sensor data on the steel sleeve is specifically the physical response of the steel sleeve body during the rolling process; the rolling mill process parameters are specifically the control and process parameters during the operation of the rolling mill. S2: Based on the time-series data records, the deformation modes are distinguished, and a deformation mode feature library is established; wherein, the deformation mode feature library includes 5 deformation modes: saddle-shaped deformation mode, conical deformation mode, spiral deformation mode, local concave deformation mode, and uniform deformation mode. The feature library of deformation patterns is established using a clustering analysis and inductive approach, specifically as follows: Establish threshold conditions for process parameters of several of the aforementioned deformation modes; Time series data records are classified based on process parameter threshold triggering conditions; Centroids are calculated for time-series data records in each category, and clustering is performed based on the centroids. The thresholds in the deformation pattern feature library are as follows: The threshold conditions for the saddle-shaped deformation mode are: the main rolling force is greater than 18000kN and the strip thickness is less than 1.5mm; The threshold conditions for the conical deformation mode are: the strip tension setting value is greater than 5kN and the difference in radial deformation at the ends is greater than 0.1mm; The threshold conditions for the spiral deformation mode are: real-time rolling speed greater than 0.5 m / s and end radial deformation change rate greater than 0.05 mm / s; The threshold condition for the local indentation deformation mode is: cooling water flow rate greater than 50 m³ / h. 3 / h / s and the peak value of the radial deformation at the end is greater than 0.3mm; Uniform deformation mode threshold condition: triggered when real-time parameters do not meet the above threshold; S3: Extract real-time process parameters from time-series data records, compare them with the deformation mode feature library, and identify deformation modes that conform to the current working conditions. S4: Assess the risk level based on the deformation mode under the current working conditions; The rolling mill steel sleeve with electronic identification chip includes a sleeve body. The sleeve body has blind mounting holes on both end faces. A smart sensor chip module is fixedly sealed within each blind mounting hole. The smart sensor chip module is an integrated microelectronic system, internally encapsulated with: The identity recognition unit is used to permanently store the unique identity and attribute information of the steel sleeve; The deformation sensing unit is used to measure the local deformation and vibration state of the cylinder end face under working load in real time. The data processing and storage unit is used to process, calculate and store the signals collected by the deformation sensing unit; A wireless transceiver unit is used for wireless data exchange with external reading and writing devices; and An energy harvesting unit is used to harvest the vibration energy of the steel sleeve during operation and convert it into electrical energy; The identity recognition unit, deformation sensing unit, data processing and storage unit, and wireless transceiver unit are electrically connected in sequence, and the energy harvesting unit provides working power to the other units in the intelligent sensing chip module.

2. The data information management method for a rolling mill steel sleeve with an electronic identification chip according to claim 1, characterized in that: The sensing data in S1 includes the unique identifier of the steel sleeve and the radial deformation at the end. The radial deformation at the end includes the first end deformation and the second end deformation. The difference between the radial deformation at the end is the difference between the first end deformation and the second end deformation. The mill process parameters in S1 include: real-time rolling speed, strip tension setpoint, strip target thickness, main rolling force, and cooling water flow rate; The integration method is an interpolation algorithm based on a sliding time window.

3. The data information management method for a rolling mill steel sleeve with an electronic identification chip according to claim 1, characterized in that: The comparison process between S3 and the deformation pattern feature library includes threshold screening and weighted Euclidean distance precise matching. The specific steps are as follows: Compare real-time process parameters with parameter thresholds; Determine if it matches the parameter threshold; Based on the judgment result, determine whether to perform weighted Euclidean distance exact matching.

4. The data information management method for a rolling mill steel sleeve with an electronic identification chip according to claim 3, characterized in that: The criterion for determining the weighted Euclidean distance exact matching is as follows: If the parameter threshold matches any one or more of the saddle-shaped deformation mode, cone-shaped deformation mode, spiral deformation mode, and local concave deformation mode, then a weighted Euclidean distance is used for precise matching. Otherwise, it is determined to be a uniform deformation mode.

5. The data information management method for a rolling mill steel sleeve with an electronic identification chip according to claim 3, characterized in that: The weighted Euclidean distance precise matching specifically involves quantizing and comparing the real-time process parameter vector with the parameterized trigger condition set of all modes.

6. The data information management method for a rolling mill steel sleeve with an electronic identification chip according to claim 1, characterized in that: The risk levels in S4 include high risk, medium risk, and low risk, and the deformation modes of the current working condition correspond to the respective risk levels; wherein, The high-risk pattern is the saddle-shaped deformation mode; Medium-risk conditions include conical deformation mode, spiral deformation mode, and localized indentation deformation; Low risk is the uniform deformation mode.

Images