A calendering process for an aluminum foil material with uniform thickness

Abstract

The present application relates to the technical field of aluminum metal rolling, in particular to a calendering process of aluminum foil material with uniform thickness, comprising: obtaining roller parameter data of a calender in the aluminum foil rolling process and thickness data of aluminum material before and after rolling; performing local uniformity analysis on the thickness data to divide uneven sections, determining thickness variation mode of the uneven sections according to data point distribution characteristics in the uneven sections, and obtaining correlation weight between the thickness variation mode of the uneven sections and each roller parameter in combination with the roller parameter data of the uneven sections in the rolling process; calculating deviation degree of the uneven sections of the current aluminum material by using the difference between the thickness variation mode of the uneven sections of the current aluminum material and historical aluminum material, and regulating the roller parameters in combination with the correlation weight, and rolling the current aluminum material by using the regulated roller parameters in the next rolling process. The present application improves the uniformity of the thickness of the aluminum foil material after calendering.

Description

Technical Field

[0001] This invention relates to the field of aluminum metal rolling technology, and more specifically to a rolling process for aluminum foil materials with uniform thickness. Background Technology

[0002] Aluminum foil is widely used in packaging, electronics, and construction, and the uniformity of aluminum foil thickness is an important quality indicator. Aluminum foil with excellent uniformity can bring enterprises a more competitive position in the market and generate huge economic benefits.

[0003] In the aluminum foil rolling process, the pressure applied by the rolls in the calender during the rolling process is an important parameter affecting the uniformity of the aluminum foil. Since the thickness of aluminum foil is usually only a few micrometers, it is very thin. Therefore, if the pressure is too high during the rolling process, the aluminum foil will be crushed or torn, and the expected thickness and quality requirements cannot be achieved. If the pressure is too low, the aluminum foil cannot be fully stretched, resulting in uneven thickness of the aluminum foil. Summary of the Invention

[0004] This invention provides a rolling process for aluminum foil materials with uniform thickness to solve existing problems.

[0005] The present invention provides a rolling process for aluminum foil materials with uniform thickness, employing the following technical solution:

[0006] One embodiment of the present invention provides a rolling process for aluminum foil material with uniform thickness, the process comprising the following steps:

[0007] Obtain roller parameter data of the rolling mill during the aluminum foil rolling process, as well as the thickness data of the aluminum material before and after rolling.

[0008] Local uniformity analysis is performed on the thickness data to divide it into several non-uniform segments. The thickness variation pattern of the non-uniform segments is determined based on the distribution characteristics of data points in the non-uniform segments. Combined with the roller parameter data of the non-uniform segments during the rolling process, the correlation weight between the thickness variation pattern of the non-uniform segments and various roller parameters is obtained.

[0009] By utilizing the differences in the thickness variation patterns of the non-uniform segments of the current aluminum material and historical aluminum materials, the deviation degree of the non-uniform segments of the current aluminum material is calculated. Combined with the aforementioned correlation weights, the roller parameters are adjusted to obtain the adjusted roller parameters.

[0010] In the next rolling process, the current aluminum material is rolled using the adjusted roller parameters.

[0011] Optionally, the step of performing local uniformity analysis on the thickness data, dividing it into several non-uniform segments, determining the thickness variation pattern of the non-uniform segments based on the distribution characteristics of data points in the non-uniform segments, and obtaining the correlation weight between the thickness variation pattern of the non-uniform segments and various roll parameters by combining the roll parameter data of the non-uniform segments during the rolling process, includes the following specific methods:

[0012] Local uniformity analysis is performed on arbitrary post-rolling thickness data to obtain several non-uniform segments in the corresponding thickness data.

[0013] Obtain the non-uniform segments from the post-rolling thickness data of all historical aluminum materials, and analyze the thickness variation pattern of each non-uniform segment based on the thickness variation reflected by the values ​​of the data points in the non-uniform segments.

[0014] By utilizing the correlation between non-uniform segments under all the same thickness variation patterns and historical roller parameter data, the correlation weight between each thickness variation pattern and each roller parameter is determined.

[0015] Optionally, the specific method for performing local uniformity analysis on arbitrary post-rolling thickness data to obtain several non-uniform segments in the corresponding thickness data includes:

[0016] The data points in the post-rolling thickness data are traversed at each time step, and a neighborhood radius is preset. Based on the numerical distribution of the data points in the post-rolling thickness data within the neighborhood radius, the local uniformity of the data points in the post-rolling thickness data is calculated.

[0017] The DBSCAN clustering algorithm is used to cluster the data points in the post-rolling thickness data. During the clustering process, the absolute value of the difference in local uniformity between adjacent data points in the post-rolling thickness data is used as the distance metric of the DBSCAN clustering algorithm, thereby obtaining several clusters, which are denoted as thickness data clusters. The average value of the local uniformity of all data points in any thickness data cluster is obtained as the uniformity level value of the thickness data cluster.

[0018] A preset uniformity threshold is set, and data segments formed by data points contained in thickness data clusters with uniformity levels less than or equal to the uniformity threshold are defined as non-uniform segments.

[0019] Optionally, the specific method for obtaining non-uniform segments from all historical aluminum material post-rolling thickness data and analyzing the thickness variation pattern of each non-uniform segment based on the thickness variation reflected by the data points in the non-uniform segments includes:

[0020] Create an empty high-dimensional array and obtain the number of data points contained in the uneven segment, denoted as the first parameter. ; Obtain the standard deviation of the thickness values ​​corresponding to all data points contained in the non-uniform segment, and denote it as the second parameter. Based on the relative position of the non-uniform segment in the post-rolling thickness data, and the relationship between the post-rolling thickness data and the pre-rolling thickness data under the corresponding rolling process, the third parameter of the non-uniform segment is calculated. The first parameter, the second parameter, and the third parameter are placed into the high-dimensional array to serve as the feature array of the non-uniform segment.

[0021] The Euclidean distance between feature arrays is used as the distance metric for the DBSCAN clustering algorithm. The DBSCAN clustering algorithm is then used to cluster the non-uniform segments of all historical aluminum materials under all rolling processes to obtain several non-uniform segment clusters. Each non-uniform segment cluster is then used as a thickness variation pattern.

[0022] Optionally, the specific method for obtaining the third parameter of the non-uniform segment is as follows:

[0023] The ratio between the number of data points contained in the non-uniform segment and the number of data points contained in the post-rolling thickness data is recorded as the first ratio. The average of the time corresponding to the last data point of the non-uniform segment and the time corresponding to the first data point is recorded as the center time of the non-uniform segment. The average of the time corresponding to the last data point and the time corresponding to the first data point in the post-rolling thickness data of the non-uniform segment is recorded as the center time of the post-rolling thickness data of the non-uniform segment. The ratio between the center time of the non-uniform segment and the center time of the post-rolling thickness data of the non-uniform segment is recorded as the second ratio. The ratio between the number of data points contained in the pre-rolling thickness data of the rolling process corresponding to the post-rolling thickness data of the non-uniform segment and the number of data points contained in the post-rolling thickness data of the non-uniform segment is recorded as the third ratio. Based on the first ratio, the second ratio, and the third ratio, the third parameter of the non-uniform segment is obtained, wherein the first ratio, the second ratio, and the third ratio are all positively correlated with the corresponding third parameter.

[0024] Optionally, the method for calculating the deviation of the non-uniform segment of the current aluminum material by utilizing the difference in thickness variation patterns between the current and historical aluminum materials, and then adjusting the roller parameters based on the associated weights to obtain the adjusted roller parameters, includes the following specific methods:

[0025] Obtain the non-uniform segment of the current aluminum material, and determine the thickness variation pattern and deviation of the non-uniform segment based on the feature array of the non-uniform segment.

[0026] The correlation weight between the thickness variation pattern corresponding to the non-uniform segment of the current interest rate material and the roller parameters is obtained, and the roller parameters are adjusted in combination with the deviation of the thickness variation pattern to obtain the adjusted roller parameters under the non-uniform segment.

[0027] Optionally, the specific method for obtaining the non-uniform segment of the current aluminum material and determining the thickness variation pattern and deviation of the non-uniform segment based on the feature array of the non-uniform segment of the current aluminum material includes:

[0028] Obtain the feature array of any non-uniform segment of the current aluminum material, denoted as the current feature array. Participate the current feature array in the clustering process of obtaining the thickness variation pattern, obtain the non-uniform segment cluster to which the current feature array belongs, and use it as the thickness variation pattern of the non-uniform segment corresponding to the current feature array. Calculate the deviation degree of the non-uniform segment corresponding to the current feature array based on the spatial distribution characteristics of the current feature array in the non-uniform segment cluster.

[0029] Optionally, the specific method for calculating the deviation degree of the non-uniform segment corresponding to the current feature array based on the spatial distribution characteristics of the current feature array in the non-uniform segment cluster includes:

[0030] This section focuses on the core aspects of the DBSCAN clustering algorithm's clustering process for feature arrays of uneven segments.

[0031] The distance ratio is calculated as follows: the ratio of the Euclidean distance between the feature array of the non-uniform segment and the cluster center of the thickness variation pattern to the average distance from all core points to the cluster center; the density ratio is calculated as follows: the ratio of the Gaussian kernel density estimate of the feature array of the non-uniform segment in the non-uniform segment cluster to the average Gaussian kernel density estimate of all core points; the deviation is calculated based on the distance ratio and the density ratio, where the distance ratio is positively correlated with the deviation and the density ratio is negatively correlated with the deviation.

[0032] Optionally, the specific method for obtaining the correlation weight between the thickness variation pattern corresponding to the non-uniform segment of the current interest rate material and the roller parameters, and adjusting the roller parameters in combination with the deviation of the thickness variation pattern to obtain the adjusted roller parameters under the non-uniform segment, includes:

[0033] Based on the similarity between the current pre-rolling thickness data of aluminum materials and the historical pre-rolling thickness data of aluminum materials, the reference non-uniform segment of the current pre-rolling thickness data of aluminum materials is obtained in the historical pre-rolling thickness data of aluminum materials.

[0034] Obtain the thickness variation pattern corresponding to the reference non-uniform segment, and the correlation weight between the thickness variation pattern and various roller parameters.

[0035] By combining the similarity between the current non-uniform segment and the reference non-uniform segment, the deviation of the current non-uniform segment, and the correlation weight between the thickness change pattern of the reference non-uniform segment and various roll parameters, the roll parameters of the reference non-uniform segment under the rolling process are adjusted to obtain the adjusted roll parameters under the current non-uniform segment.

[0036] Optionally, the specific method for obtaining the reference non-uniform segment is as follows:

[0037] The DTW distance between the current pre-rolling thickness data of aluminum material and the pre-rolling thickness data of all historical aluminum materials is calculated using the Dynamic Time Warping algorithm and denoted as the first distance. and will As the first similarity, the historical pre-rolling thickness data of aluminum material corresponding to the highest first similarity is selected as the reference pre-rolling thickness data for the current aluminum material. Let be a natural constant; any non-uniform segment in the current pre-rolling thickness data of the aluminum material is denoted as the current non-uniform segment, and the absolute value of the difference between the center time of the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness data is obtained. ,Will The second similarity between the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness data is used as the basis for determining the reference non-uniform segment. The non-uniform segment in the reference pre-rolling thickness data corresponding to the maximum second similarity is selected as the reference non-uniform segment for the current non-uniform segment. This represents the DTW distance between the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness layout. This represents an exponential function with the natural constant as its base.

[0038] The beneficial effects of the technical solution of the present invention are as follows: by combining local uniformity analysis, non-uniform segments of aluminum foil are identified and thickness change patterns are extracted. Based on historical data, a correlation weight model between each pattern and parameters such as roll speed, roll gap, and speed difference is established. Combined with deviation degree and similarity, adaptive control of the current rolling state is achieved, which improves the control accuracy of aluminum foil thickness uniformity, effectively avoids damage and unevenness caused by over-pressure or under-pressure, and improves the thickness uniformity of aluminum foil. Attached Figure Description

[0039] 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart illustrating the steps of a uniform thickness aluminum foil rolling process according to the present invention.

[0041] Figure 2 This is a flowchart illustrating the steps for obtaining association weight features according to an embodiment of the present invention.

[0042] Figure 3This is a flowchart illustrating the steps for obtaining the adjusted roller parameters according to an embodiment of the present invention. Detailed Implementation

[0043] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a uniform thickness aluminum foil rolling process proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0045] The following describes in detail, with reference to the accompanying drawings, a specific scheme for the rolling process of aluminum foil material with uniform thickness provided by the present invention.

[0046] Please see Figure 1 The diagram illustrates a process flow chart of a uniform thickness aluminum foil rolling process according to an embodiment of the present invention, which includes the following steps:

[0047] Step S001: Obtain the roller parameter data of the rolling mill during the aluminum foil rolling process and the thickness data of the aluminum material before and after rolling.

[0048] It should be noted that the rolling mill used for aluminum foil rolling typically includes a frame, rollers, a gap adjustment device, and a roller deflection compensation device. The rollers are usually arranged symmetrically and apply pressure from above and below to cause the aluminum material to undergo plastic deformation. In the specific rolling process, aluminum ingots or sheets are first fed into the gap between the rollers. The rollers rotate under the drive of the motor and apply pressure to cause the aluminum material to undergo plastic flow, reducing its thickness and extending its length. Then, through multiple rolling processes, the aluminum material is rolled to be thinner and more uniform until the material is rolled to the target thickness. Subsequently, it is shaped by a cooling system. During this process, it is necessary to accurately control the pressure applied by the rollers to the aluminum ingots or sheets to ensure uniform thickness and achieve the target thickness.

[0049] Specifically, in order to achieve the uniform thickness aluminum foil rolling process proposed in this embodiment, it is first necessary to collect the roller parameter data of the rolling mill during the aluminum foil rolling process and the thickness data of the aluminum material after each rolling. The specific process is as follows:

[0050] First, a fixed sampling frequency is used to acquire the thickness data of the aluminum material before and after each rolling process using a thickness measuring device. These are recorded as pre-roll thickness data and post-roll thickness data, and are collectively referred to as thickness data. The time sequence data of the roller parameters formed during each rolling of the aluminum material is recorded and recorded as roller parameter data. The roller parameter data includes: average roller speed data, speed difference data, and roller gap data.

[0051] As an optional embodiment, the specific method for obtaining the thickness data and roll parameter data is as follows: a fixed sampling frequency is preset, and a non-contact thickness gauge (such as an X-ray thickness gauge, a beta-ray thickness gauge, or a laser rangefinder) is used to monitor the thickness value of the aluminum material when it passes through the roll gap before and after each rolling pass in real time; process parameters related to the roll operating status are collected in real time through the encoder, displacement sensor, and PLC module integrated in the rolling mill control system.

[0052] It should be noted that, in the embodiments of the present invention, the fixed sampling frequency is preset to 10Hz based on experience. The specific value depends on the process response speed, and the embodiments of the present invention do not impose specific limitations.

[0053] It should be noted that for the roller parameter data, the roller speed data reflects the average rotational speed of the upper and lower rollers at each moment, the speed difference data reflects the speed difference between the upper and lower rollers at each moment, and the roller gap data reflects the distance between the upper and lower rollers at each moment.

[0054] Then, the thickness data of the aluminum material and the roller parameter data are obtained when the aluminum material is rolled several times in the historical rolling process, and are recorded as historical thickness data and historical roller parameter data respectively. The aluminum material in the historical rolling process is called historical aluminum material. The aluminum material processed in the historical rolling process is used as historical aluminum material, and the aluminum material processed in the current rolling process is used as current aluminum material.

[0055] It should be noted that, since aluminum materials need to undergo multiple rolling processes in the rollers to achieve the target thickness during the rolling process, in order to comprehensively understand the thickness changes of aluminum materials and roller parameters throughout the entire rolling process, this embodiment of the invention selects to acquire the thickness data and roller parameter data of aluminum materials during the rolling process, and the thickness data and roller parameter data are time-series data; in addition, since any aluminum material usually needs to undergo several rolling processes to achieve the target thickness, the thickness data and roller parameter data acquired in this embodiment of the invention are generated for each rolling process.

[0056] Finally, the thickness data and roller parameter data of any aluminum material in any rolling process are taken as a set of data under the rolling process, called the rolling data set. Then, one rolling process corresponds to one rolling data set. In addition, the target thickness set under each rolling process is obtained.

[0057] Thus, the roller parameter data and thickness data are obtained through the above method.

[0058] Step S002: Perform local uniformity analysis on the thickness data, divide it into several non-uniform segments, determine the thickness variation pattern of the non-uniform segments based on the distribution characteristics of data points in the non-uniform segments, and obtain the correlation weight between the thickness variation pattern of the non-uniform segments and various roller parameters by combining the roller parameter data of the non-uniform segments during the rolling process.

[0059] It should be noted that all the data included in the roller parameter data affect the pressure generated during the rolling process of aluminum material in different ways. However, since the specific effects of the generated pressure on the aluminum material are different, the thickness change of the aluminum material after each rolling is different in pattern. Therefore, in order to facilitate accurate adjustment of the roller parameters in the future, this embodiment of the invention selects to analyze the historical thickness data of aluminum material and roller parameter data to determine the correspondence between different thickness change patterns and roller parameters.

[0060] Specifically, in step S201, a local uniformity analysis is performed on any post-rolling thickness data to obtain several non-uniform segments in the corresponding thickness data.

[0061] As a preferred embodiment, the method for obtaining the non-uniform segment includes:

[0062] First, the data points in the post-rolling thickness data are traversed time-by-time, and a neighborhood radius is preset. Based on the numerical distribution of the data points in the post-rolling thickness data within the neighborhood radius, the local uniformity of the data points in the post-rolling thickness data is calculated.

[0063] Then, the DBSCAN clustering algorithm is used to cluster the data points in the post-rolling thickness data. During the clustering process, the absolute value of the difference in local uniformity between adjacent data points in the post-rolling thickness data is used as the distance metric of the DBSCAN clustering algorithm, thereby obtaining several clusters, which are denoted as thickness data clusters. The average value of the local uniformity of all data points in any thickness data cluster is obtained as the uniformity level value of the thickness data cluster.

[0064] Finally, a uniformity threshold is preset, and data segments formed by data points contained in thickness data clusters with uniformity levels less than or equal to the uniformity threshold are defined as non-uniform segments.

[0065] It should be noted that, in the embodiments of the present invention, the neighborhood radius is preset to 2 based on experience, which can be adjusted according to the actual situation. The embodiments of the present invention do not impose specific limitations. In addition, the DBSCAN clustering algorithm is an existing clustering algorithm, so it will not be described in detail in the embodiments of the present invention.

[0066] As an optional embodiment, the specific method for calculating the local homogeneity is as follows:

[0067]

[0068] in, Indicates the first [thickness] in the post-rolling thickness data Local uniformity of data points; Indicates the thickness data after rolling. The standard deviation of the thickness values ​​of all data points within the neighborhood radius of each data point; Indicates the preset neighborhood radius; Indicates the thickness data after rolling. Within the neighborhood radius of the data point, the first Thickness value of each data point; This indicates the target thickness during the rolling process corresponding to the post-rolling thickness data; Represents the absolute value function; This represents an exponential function with the natural constant as its base.

[0069] It should be noted that the local uniformity is used to describe the uniformity of the thickness distribution of the corresponding data point within the neighborhood radius. The larger the local uniformity value, the higher the degree of non-uniformity of the aluminum material thickness within the neighborhood radius at the corresponding position after rolling.

[0070] Step S202: Obtain the non-uniform segments in the post-rolling thickness data of all historical aluminum materials, and analyze the thickness change pattern of each non-uniform segment based on the thickness change reflected by the values ​​of the data points in the non-uniform segments.

[0071] It should be noted that roll gap, roll speed, and speed difference are the main adjustment methods for controlling the thickness of aluminum foil. However, they have fundamentally different ways of affecting the thickness of the rolled aluminum foil and their ability to do so. Roll gap directly determines the amount of reduction of the aluminum material, but in the actual deformation process, due to roll deflection and elastic deformation, the pressure distribution is not uniform, resulting in different thickness change patterns at the edges and in the center. When the aluminum foil thickness is small, the rolling mill usually operates in rolling force mode, where even small parameter changes can lead to significant thickness fluctuations. Roll speed affects the deformation rate and temperature field of the material in the roll gap. The rolling process is a forced deformation process of metal. Changes in the thickness and hardness of the rolled piece directly lead to dynamic changes in the rolling force. During high-speed rolling, the heat generated by material deformation cannot be dissipated in time, leading to local temperature rise, reducing the yield strength of the material, and making deformation more likely to occur. In addition, speed difference (i.e., roll speed difference) creates asymmetric deformation conditions. In asynchronous rolling, changes in speed difference directly affect the deformation distribution on both sides of the material, thereby changing the thickness distribution pattern. Therefore, in order to facilitate the effective adjustment of the roller parameters for different thickness distribution variations and thus improve the thickness uniformity of the rolled aluminum foil, this embodiment of the invention selects to analyze the non-uniform segments in the historical aluminum material post-rolling thickness data to determine the specific thickness variations of the aluminum foil in the non-uniform segments.

[0072] As a preferred embodiment, the method for obtaining the thickness variation pattern includes:

[0073] First, feature extraction is performed on any uneven segment to obtain a feature array of the uneven segment, which is used to describe the specific data distribution of the uneven segment.

[0074] As an optional embodiment, for any non-uniform segment, the specific method for obtaining the feature array of the non-uniform segment is as follows: an empty high-dimensional array is established, and the number of data points contained in the non-uniform segment is obtained, denoted as the first parameter. ; Obtain the standard deviation of the thickness values ​​corresponding to all data points contained in the non-uniform segment, and denote it as the second parameter. Based on the relative position of the non-uniform segment in the post-rolling thickness data, and the relationship between the post-rolling thickness data and the pre-rolling thickness data under the corresponding rolling process, the third parameter of the non-uniform segment is calculated. The first parameter, the second parameter, and the third parameter are placed into the high-dimensional array to serve as the feature array of the non-uniform segment.

[0075] As a preferred embodiment, the method for obtaining the third parameter of the non-uniform segment is as follows: The ratio between the number of data points contained in the non-uniform segment and the number of data points contained in the post-rolling thickness data is recorded as a first ratio; the average of the time corresponding to the last data point of the non-uniform segment and the time corresponding to the first data point is obtained as the center time of the non-uniform segment; the average of the time corresponding to the last data point and the time corresponding to the first data point in the post-rolling thickness data to which the non-uniform segment belongs is obtained as the center time of the post-rolling thickness data to which the non-uniform segment belongs; the ratio between the center time of the non-uniform segment and the center time of the post-rolling thickness data to which the non-uniform segment belongs is recorded as a second ratio; the ratio between the number of data points contained in the pre-rolling thickness data corresponding to the post-rolling thickness data of the non-uniform segment and the number of data points contained in the post-rolling thickness data to which the non-uniform segment belongs is recorded as a third ratio; the third parameter of the non-uniform segment is obtained based on the first ratio, the second ratio, and the third ratio, wherein the first ratio, the second ratio, and the third ratio are all positively correlated with the corresponding third parameter.

[0076] As an optional embodiment, for any non-uniform segment, the specific calculation method for the third parameter of the non-uniform segment is as follows:

[0077]

[0078] in, The third parameter represents the non-uniform segment; This indicates the number of data points contained in the uneven segment. This indicates the number of data points included in the post-rolling thickness data of the non-uniform segment; Indicates the center time of the non-uniform segment; This indicates the center time of the post-rolling thickness data to which the non-uniform segment belongs; This indicates the number of data points contained in the pre-rolling thickness data of the rolling process corresponding to the post-rolling thickness data of the non-uniform segment. This indicates the time corresponding to the last data point in the uneven segment; This indicates the time corresponding to the first data point in the uneven segment; This indicates the time corresponding to the last data point in the post-rolling thickness data of the non-uniform segment; This indicates the time corresponding to the first data point in the post-rolling thickness data of the non-uniform segment.

[0079] It should be noted that the third parameter is used to reflect the comprehensive characteristic quantity of the relative position and relative scale of the non-uniform segment, where This indicates the proportion of the duration of the non-uniformity phenomenon to the entire rolling process. This indicates the point in time when the non-uniformity occurred relative to the entire rolling process; This represents the ratio of the number of data points before and after rolling, reflecting the material elongation. The smaller the value, the greater the material elongation, meaning the material is rolled thinner and longer.

[0080] Then, clustering algorithms are used in conjunction with feature arrays to divide the non-uniform segments, resulting in several thickness variation patterns.

[0081] As an optional embodiment, the specific method for obtaining the thickness variation pattern is as follows: the Euclidean distance between feature arrays is used as the distance metric of the DBSCAN clustering algorithm, and the DBSCAN clustering algorithm is used to cluster the non-uniform segments of all historical aluminum materials under all rolling processes to obtain several non-uniform segment clusters, and a non-uniform segment cluster is taken as a thickness variation pattern.

[0082] It should be noted that by dividing the post-rolling thickness data of historical aluminum materials into non-uniform segments and combining the specific manifestations of the non-uniform segments of all historical aluminum materials, the thickness variation patterns exhibited by the thickness data in the non-uniform segments can be analyzed. This facilitates the subsequent determination of the relevant roller parameters and specific related situations through each thickness variation pattern. This allows for the adjustment of roller parameters using these related situations, thereby effectively controlling the working process of the rollers in the aluminum foil rolling process and improving the uniformity of aluminum foil thickness.

[0083] Step S203: Using the correlation between the non-uniform segments under all the same thickness variation patterns and historical roller parameter data, determine the correlation weight between each thickness variation pattern and each roller parameter.

[0084] It should be noted that during the rolling process of aluminum, the rolling pressure exerted by the rollers on the aluminum is affected by a variety of roller parameters, and the degree of influence of each roller parameter varies. Therefore, the unevenness is caused by the rollers forming corresponding rolling patterns on the aluminum under roller parameters with different degrees of influence. So, when adjusting the roller parameters for unevenness in aluminum under different thickness variation patterns, the correlation between the corresponding thickness variation pattern and the roller parameters can be considered.

[0085] As a preferred embodiment, the specific calculation method for the association weight is as follows:

[0086] For any thickness variation pattern, the correlation weight of the thickness variation pattern is calculated based on the correlation between the non-uniform segments included in the thickness variation pattern and the roll parameters under the corresponding rolling process.

[0087] As an optional embodiment, the specific calculation method for the association weight is as follows:

[0088]

[0089] in, Indicates the first The thickness variation pattern and the first The correlation weights between the parameters of each roller; Indicates the first The total number of non-uniform segments contained in each thickness variation pattern; Indicates the first In the thickness variation mode, the first The first feature array of the i-th uneven segment One component; Indicates the first The component in the first The mean value among the thickness variation patterns; Indicates the first The component in the first Standard deviation in each thickness variation pattern; Indicates the first In the thickness variation mode, the first The post-rolling thickness data of the non-uniform segment corresponds to the first rolling process. Individual roller parameter values; Indicates the first The roller parameter at the first The mean value among the thickness variation patterns; Indicates the first The roller parameter at the first Standard deviation in each thickness variation pattern; This represents the absolute value function.

[0090] It should be noted that the correlation weight reflects the degree of correlation between the corresponding thickness variation pattern and the roller parameters. That is, when the rolled aluminum material exhibits uneven thickness variation, the extent to which this uneven thickness variation pattern is affected by the corresponding roller parameters is considered. The larger the correlation weight value, the greater the influence of the roller parameters on the thickness variation pattern of the rolled aluminum material. In the formula for calculating the correlation weight, ... and The product of the characteristic components and the roller parameters represents the covariance between the two, reflecting their correlation.

[0091] Thus, several thickness variation modes and their correlation weights with roller parameters were obtained using the above method, such as... Figure 2 The diagram shown is a flowchart of the steps for obtaining association weights in an embodiment of the present invention.

[0092] Step S003: Calculate the deviation of the non-uniform segment of the current aluminum material by utilizing the difference in the thickness variation pattern of the non-uniform segment between the current aluminum material and the historical aluminum material, and adjust the roller parameters in combination with the aforementioned correlation weights to obtain the adjusted roller parameters.

[0093] It should be noted that, under normal circumstances, the roller parameters do not change significantly during the rolling of aluminum. However, in actual rolling processes, various factors can affect the thickness variation pattern of the aluminum material when uneven sections appear. This pattern may not strictly match the thickness variation pattern observed in historical uneven sections, and there is a certain deviation. Therefore, when using the correlation weight between the thickness variation pattern and roller parameters to control the roller parameters during the rolling process of the current aluminum material, the specific deviation between the thickness variation pattern of the uneven section in the current aluminum material and the thickness variation pattern obtained from historical aluminum materials should also be considered, and the process of controlling the roller parameters using the correlation weight should be corrected.

[0094] Specifically, in step S301, the non-uniform segment of the current aluminum material is obtained, and the thickness variation pattern and deviation of the non-uniform segment are determined based on the feature array of the non-uniform segment of the current aluminum material.

[0095] As a preferred embodiment, the method for obtaining the deviation is as follows:

[0096] Obtain the feature array of any non-uniform segment of the current aluminum material, denoted as the current feature array. Participate the current feature array in the clustering process of obtaining the thickness variation pattern, obtain the non-uniform segment cluster to which the current feature array belongs, and use it as the thickness variation pattern of the non-uniform segment corresponding to the current feature array. Calculate the deviation degree of the non-uniform segment corresponding to the current feature array based on the spatial distribution characteristics of the current feature array in the non-uniform segment cluster.

[0097] As an optional embodiment, for any non-uniform segment of the current aluminum material, the specific calculation method for the deviation of the non-uniform segment is as follows:

[0098] First, we will identify the core aspects of the DBSCAN clustering algorithm's clustering process for the feature array of uneven segments.

[0099] Then, the Euclidean distance between the feature array of the non-uniform segment and the cluster center of the thickness variation pattern is obtained, and the ratio of this distance to the average distance from all core points to the cluster center is recorded as the distance ratio. The Gaussian kernel density estimate of the feature array of the non-uniform segment in the non-uniform segment cluster is obtained, and the ratio of this estimate to the average Gaussian kernel density estimate of all core points is recorded as the density ratio. Based on the distance ratio and the density ratio, the deviation is obtained, wherein the distance ratio is positively correlated with the deviation and the density ratio is negatively correlated with the deviation.

[0100] As an optional embodiment, the specific method for calculating the deviation is as follows:

[0101]

[0102] in, This indicates the degree of deviation of the non-uniform segment of the current aluminum material; The Euclidean distance between the feature array of the non-uniform segment and the cluster center of the thickness variation pattern to which it belongs; This represents the average distance from all core points to the cluster center; This represents the Gaussian kernel density estimate of the feature array of the non-uniform segment in its respective non-uniform segment cluster; This represents the average Gaussian kernel density estimate for all core points.

[0103] It should be noted that the deviation degree is used to describe the thickness change pattern exhibited by the thickness value in the corresponding non-uniform segment, and the degree of deviation from the thickness change pattern formed by the feature array corresponding to the non-uniform segment of historical aluminum material. The larger the deviation degree value, the more it indicates that although the thickness change pattern of the corresponding non-uniform segment exists in the cluster of non-uniform segments of historical aluminum material, there is still a certain deviation overall. Therefore, directly using the correlation parameter between the thickness change pattern of historical aluminum material and the roller parameter to adjust the roller parameter will lead to the problem of excessive adjustment amount, thus failing to effectively ensure that the uniformity of the rolled aluminum foil meets the process requirements. Therefore, this embodiment of the invention selects to quantify the deviation of the thickness change pattern exhibited by the non-uniform segment in the current aluminum material from the historical thickness change pattern in order to correct the subsequent roller parameter adjustment process, thereby improving the control effect of the roller and optimizing the equipment operating parameters in the aluminum foil rolling process; in the calculation formula of the deviation degree, the distance factor Density factor reflects the degree of geometric deviation between the current uneven segment and the historical typical pattern. It reflects the degree of anomaly of the current uneven segment in the feature space, that is, the sparsity of the surrounding data points.

[0104] Step S302: Obtain the correlation weight between the thickness change pattern corresponding to the non-uniform segment of the current interest rate material and the roller parameters, and adjust the roller parameters in combination with the deviation of the thickness change pattern to obtain the adjusted roller parameters under the non-uniform segment.

[0105] First, based on the similarity between the current pre-rolling thickness data of aluminum materials and the historical pre-rolling thickness data of aluminum materials, a reference non-uniform segment in the historical pre-rolling thickness data of aluminum materials is obtained for the non-uniform segment of the current pre-rolling thickness data of aluminum materials.

[0106] As an optional embodiment, the method for obtaining the reference non-uniform segment is as follows:

[0107] The DTW distance between the current pre-rolling thickness data of aluminum material and the pre-rolling thickness data of all historical aluminum materials is calculated using the Dynamic Time Warping algorithm and denoted as the first distance. and will As the first similarity, the historical pre-rolling thickness data of aluminum material corresponding to the highest first similarity is selected as the reference pre-rolling thickness data for the current aluminum material. Let be a natural constant. Any non-uniform segment in the current pre-rolling thickness data of the aluminum material is denoted as the current non-uniform segment. The absolute value of the difference between the center time of the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness data is obtained. ,Will The second similarity between the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness data is used as the basis for determining the reference non-uniform segment. The non-uniform segment in the reference pre-rolling thickness data corresponding to the maximum second similarity is selected as the reference non-uniform segment for the current non-uniform segment. This represents the DTW distance between the current non-uniform segment and any non-uniform segment in the reference pre-rolling thickness layout. This represents an exponential function with the natural constant as its base.

[0108] Then, the thickness variation pattern corresponding to the reference non-uniform segment is obtained, as well as the correlation weight between the thickness variation pattern and various roller parameters.

[0109] Finally, by combining the similarity between the current non-uniform segment and the reference non-uniform segment, the deviation of the current non-uniform segment, and the correlation weight between the thickness change pattern of the reference non-uniform segment and various roll parameters, the roll parameters of the reference non-uniform segment under the rolling process are adjusted to obtain the adjusted roll parameters under the current non-uniform segment.

[0110] As an optional embodiment, the specific calculation method for the adjusted roller parameters under the current non-uniform segment is as follows:

[0111]

[0112] in, Indicates the first (i)th (i)th segment under the current non-uniformity. Adjust the roller parameters after adjustment; This indicates the roller parameters of the reference non-uniform segment under the corresponding rolling process for the current non-uniform segment; This indicates the thickness variation pattern corresponding to the reference non-uniform segment and the first... The correlation weights between the parameters of each roller; This indicates the degree of deviation of the current non-uniform segment; This indicates the second similarity between the current non-uniform segment and the reference non-uniform segment; Indicates the first segment of the current non-uniform segment. Adjustment direction factor for each roller parameter.

[0113] As an optional embodiment, the method for obtaining the adjustment direction factor of the roller parameters is as follows: The average thickness value corresponding to all data points in the current non-uniform segment is obtained and recorded as the average thickness of the current non-uniform segment. If the average thickness of the current non-uniform segment is greater than the target thickness of the corresponding rolling stage, then the adjustment direction factor of the roller gap parameter is 1, and the adjustment direction factors of the roller speed and speed difference are -1; if the average thickness of the current non-uniform segment is greater than or equal to the target thickness of the corresponding rolling stage, then... If the average thickness of the current non-uniform section is less than the target thickness of the corresponding rolling stage, the adjustment direction factor of the roll gap parameter is -1, and the adjustment direction factor of the roll speed and speed difference is 1.

[0114] It should be noted that, for the roll gap parameter, a decrease in roll gap results in an increase in rolling pressure, and vice versa; for the roll speed parameter, an increase in roll speed significantly increases the material flow stress, thereby increasing the rolling pressure, and vice versa; for the speed difference parameter, an increase in speed difference generates greater shear stress inside the aluminum material, resulting in an increase in rolling pressure, and vice versa; therefore, in this embodiment of the invention, the adjustment direction factor is set based on the above principles.

[0115] Thus, the adjusted roller parameters are obtained through the above method, as follows: Figure 3 The diagram shown is a flowchart illustrating the steps for obtaining the adjusted roller parameters in an embodiment of the present invention.

[0116] Step S004: In the next rolling process, the current aluminum material is rolled using the adjusted roller parameters.

[0117] Specifically, firstly, the adjusted roller parameters are sent to the main control system of the calender (such as a PLC or DCS system), and the following equipment actuators are automatically configured: the roller gap between the upper and lower rollers is adjusted by the roller gap adjustment device to accurately set the roller gap opening; the rotational speed of the upper and lower rollers is adjusted by the frequency conversion drive system, and the speed difference is controlled.

[0118] Then, the current aluminum material is fed into the roll gap and rolled in the next pass under the adjusted roll parameters. During the rolling process, the tension is kept stable and the cooling system is kept running normally to ensure that the material completes plastic deformation under controlled conditions.

[0119] Finally, during this rolling process, new pre-roll thickness data, post-roll thickness data, and corresponding roll parameter data are collected at a preset sampling frequency to form a new set of rolling data, which serves as the input basis for the next round of analysis and control.

[0120] This concludes the embodiment.

[0121] It should be noted that the embodiments used in this example The model is only used to represent negative correlations and the results of the constraint model output are in Within this range, in specific implementations, other models with the same purpose can be substituted; this embodiment is merely an example. The description will be based on a model, without making specific limitations on it. This refers to the input of the model.

[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A calendering process for an aluminum foil material of uniform thickness, characterized by, The process comprises the following steps: Obtain the roll parameters of the calender during the aluminum foil rolling process and the thickness data of the aluminum material before and after rolling; Perform local uniformity analysis on the thickness data, divide into several uneven sections, determine the thickness variation mode of the uneven section according to the data point distribution characteristics in the uneven section, and obtain the correlation weight between the thickness variation mode of the uneven section and each roll parameter in combination with the roll parameter data of the uneven section during the rolling process; Calculate the deviation degree of the uneven section of the current aluminum material by using the difference in the thickness variation mode of the uneven section of the current aluminum material and the historical aluminum material, and combine the correlation weight to regulate the roll parameters to obtain the regulated roll parameters; In the next rolling process, the current aluminum material is rolled by using the regulated roll parameters.

2. The calendering process of a uniform thickness aluminum foil material according to claim 1, wherein, The specific method for performing local uniformity analysis on the thickness data, dividing into several uneven sections, determining the thickness variation mode of the uneven section according to the data point distribution characteristics in the uneven section, and obtaining the correlation weight between the thickness variation mode of the uneven section and each roll parameter in combination with the roll parameter data of the uneven section during the rolling process comprises: Perform local uniformity analysis on any post-rolling thickness data to obtain several uneven sections in the corresponding thickness data; Obtain the uneven sections in the post-rolling thickness data of all historical aluminum materials, and analyze the thickness variation mode of each uneven section according to the thickness variation reflected by the numerical values of the data points in the uneven section; Determine the correlation weight between each thickness variation mode and each roll parameter by using the correlation between the uneven sections with the same thickness variation mode and the historical roll parameter data.

3. The calendering process of a uniform thickness aluminum foil material according to claim 2, wherein, The specific method for performing local uniformity analysis on any post-rolling thickness data to obtain several uneven sections in the corresponding thickness data comprises: Iterate through the data points in the post-rolling thickness data at each time, preset a neighborhood radius, and calculate the local uniformity of the data points in the post-rolling thickness data in combination with the numerical distribution of the data points in the post-rolling thickness data within the neighborhood radius; Use the DBSCAN clustering algorithm to cluster the data points in the post-rolling thickness data, and use the absolute value of the difference between the local uniformities of adjacent data points in the post-rolling thickness data as the distance measurement method of the DBSCAN clustering algorithm during clustering to obtain several clustering clusters, denoted as thickness data clusters; obtain the average value of the local uniformity of all data points in any thickness data cluster as the uniformity level value of the thickness data cluster; Preset a uniformity threshold, and form a data section from the data points in the thickness data cluster whose uniformity level value is less than or equal to the uniformity threshold as an uneven section.

4. The calendering process of claim 2, wherein The specific method for obtaining the uneven sections in the post-rolling thickness data of all historical aluminum materials and analyzing the thickness variation mode of each uneven section according to the thickness variation reflected by the numerical values of the data points in the uneven section comprises: establishing an empty high-dimensional array, obtaining the number of data points contained in the uneven section, denoted as a first parameter ; obtaining the standard deviation of the thickness values corresponding to all data points contained in the uneven section, denoted as a second parameter ; calculating a third parameter of the uneven section according to the relative position relationship of the uneven section in the post-rolling thickness data and the change relationship between the post-rolling thickness data and the pre-rolling thickness data under the corresponding rolling process ; placing the first parameter, the second parameter and the third parameter into the high-dimensional array as the characteristic array of the uneven section; The Euclidean distance between the feature arrays is taken as the distance measurement mode of the DBSCAN clustering algorithm, and the DBSCAN clustering algorithm is used to cluster all the uneven sections of the historical aluminum materials under all rolling processes to obtain a plurality of uneven section clusters, and an uneven section cluster is taken as a thickness variation mode.

5. The calendering process of a uniform thickness aluminum foil material as claimed in claim 4, wherein, The specific method for obtaining the third parameter of the uneven section is: The ratio between the number of data points in the uneven section and the number of data points in the rolled thickness data is obtained, denoted as a first ratio; the average of the time corresponding to the last data point of the uneven section and the time corresponding to the first data point is taken as the center time of the uneven section; the average of the time corresponding to the last data point of the rolled thickness data to which the uneven section belongs and the time corresponding to the first data point is taken as the center time of the rolled thickness data to which the uneven section belongs; the ratio between the center time of the uneven section and the center time of the rolled thickness data to which the uneven section belongs is denoted as a second ratio; the ratio between the number of data points in the rolled thickness data corresponding to the rolling process of the rolled thickness data to which the uneven section belongs and the number of data points in the rolled thickness data to which the uneven section belongs is denoted as a third ratio; the third parameter of the uneven section is obtained according to the first ratio, the second ratio and the third ratio, wherein the first ratio, the second ratio and the third ratio are all positively correlated with the third parameter.

6. The calendering process of a uniform thickness aluminum foil material as claimed in claim 4, wherein, The specific method for calculating the deviation degree of the uneven section of the current aluminum material by using the difference between the thickness variation modes of the uneven sections of the current aluminum material and the historical aluminum material, and combining the correlation weight, to regulate the roll parameters to obtain the regulated roll parameters, includes: An uneven section of the current aluminum material is obtained, and the thickness variation mode and the deviation degree of the uneven section of the current aluminum material are determined according to the feature array of the uneven section of the current aluminum material. The correlation weight between the thickness variation mode corresponding to the uneven section of the current aluminum material and the roll parameters is obtained, and the roll parameters are regulated in combination with the deviation degree of the thickness variation mode to obtain the regulated roll parameters under the uneven section.

7. The calendering process of a uniform thickness aluminum foil material as claimed in claim 6, wherein, The specific method for obtaining the uneven section of the current aluminum material, and determining the thickness variation mode and the deviation degree of the uneven section of the current aluminum material according to the feature array of the uneven section of the current aluminum material, includes: The feature array of any uneven section of the current aluminum material is obtained, denoted as a current feature array, the current feature array is involved in the clustering process of obtaining the thickness variation mode, the uneven section cluster to which the current feature array belongs is obtained as the thickness variation mode to which the uneven section corresponding to the current feature array belongs, and the deviation degree of the uneven section corresponding to the current feature array is calculated according to the spatial distribution characteristics of the current feature array in the uneven section cluster.

8. The calendering process of a uniform thickness aluminum foil material as claimed in claim 7, wherein, The specific method for calculating the deviation degree of the uneven section corresponding to the current feature array according to the spatial distribution characteristics of the current feature array in the uneven section cluster, includes: The core point in the clustering process of the DBSCAN clustering algorithm on the feature array of the uneven section is obtained. The ratio of the Euclidean distance between the feature array of the uneven section and the cluster center of the thickness variation mode to the average distance between all core points is denoted as a distance ratio; the ratio of the Gaussian kernel density estimation value of the feature array of the uneven section in the uneven section cluster to the average Gaussian kernel density estimation value of all core points is denoted as a density ratio; a deviation degree is obtained according to the distance ratio and the density ratio, the distance ratio is positively correlated with the deviation degree, and the density ratio is negatively correlated with the deviation degree.

9. The calendering process of claim 6, wherein the aluminum foil material is uniformly thick. The association weight between the thickness variation mode corresponding to the uneven section of the current interest rate material and the roller parameters is obtained, and the roller parameters are regulated in combination with the deviation degree of the thickness variation mode, to obtain the regulated roller parameters under the uneven section, and the specific method comprises the following steps: According to the similarity between the pre-rolling thickness data of the current aluminum material and the pre-rolling thickness data of the historical aluminum material, a reference uneven section of the pre-rolling thickness data of the current aluminum material in the pre-rolling thickness data of the historical aluminum material is obtained; The thickness variation mode corresponding to the reference uneven section and the association weight between the thickness variation mode and each roller parameter are obtained; In combination with the similarity between the current uneven section and the reference uneven section, the deviation degree of the current uneven section, and the association weight between the thickness variation mode of the reference uneven section and each roller parameter, the roller parameters of the reference uneven section in the rolling process are regulated to obtain the regulated roller parameters under the current uneven section.

10. The calendering process of a uniform thickness aluminum foil material as claimed in claim 9, wherein, The specific method for obtaining the reference uneven section comprises the following steps: a DTW distance between the pre-rolling thickness data of the current aluminum material and the pre-rolling thickness data of all historical aluminum materials is calculated using a dynamic time warping algorithm, denoted as a first distance , and is taken as the first similarity; the pre-rolling thickness data of the historical aluminum material corresponding to the highest first similarity is selected as the reference pre-rolling thickness data of the pre-rolling thickness data of the current aluminum material, wherein is a natural constant; any uneven section in the pre-rolling thickness data of the current aluminum material is taken as a current uneven section, and a difference absolute value between a center time of the current uneven section and a center time of any uneven section in the reference pre-rolling thickness data is obtained , and is taken as a second similarity between the current uneven section and any uneven section in the reference pre-rolling thickness data; the uneven section in the reference pre-rolling thickness data corresponding to the maximum second similarity is selected as a reference uneven section of the current uneven section, wherein represents a DTW distance between the current uneven section and any uneven section in the reference pre-rolling thickness layout, represents an exponential function with a natural constant as the base number.

Images