Method, system and equipment for detecting deviation of roll gap of rolling mill

Abstract

The invention relates to the technical field of rolling technology and equipment detection, in particular to a method, a system and equipment for detecting deviation of a roll gap of a rolling mill, and the method comprises the steps: constructing a vibration detection center line parallel to a roll axis of a working roll according to a to-be-processed copper material, and according to the vibration detection center line, roller vibration characteristics are acquired by taking each working roller as a unit, a vibration characteristic matrix is generated according to the machining time of the corresponding working roller, and vibration deviation comparison is performed on the vibration characteristic matrix corresponding to each working roller by establishing a reference vibration deviation model, so that a roller gap deviation detection result is obtained. The problem that weak vibration changes under high-speed and high-load working conditions are difficult to capture in roll gap deviation is effectively solved, a reference vibration deviation model is established through dynamic monitoring and analysis based on vibration signals, efficient and accurate roll gap deviation detection is achieved, and the stability of the production process is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of rolling process and equipment detection, and in particular to a method, system and equipment for detecting deviation of a roll gap of a rolling mill. Background Art

[0002] With the continuous improvement of automation and production efficiency in the rolling industry, the operating stability and performance optimization of rolling mill equipment are of great significance to production quality, cost control and equipment maintenance. Roll gap deviation refers to the change in the distance or position between the rolls, which leads to problems such as uneven material thickness, surface defects or mechanical wear in the production process, directly affecting the quality and production efficiency of the final product. At present, the detection method of roll gap deviation mainly relies on traditional physical quantity measurement, such as monitoring of parameters such as displacement and pressure.

[0003] Although these monitoring methods can capture the macroscopic characteristics of roll gap changes to a certain extent, they have some limitations. Traditional methods often rely on manually set thresholds for judgment, and it is difficult to reflect the slight changes in roll gap deviation under complex working conditions in real time and comprehensively. At the same time, these methods lack in-depth analysis of vibration signals, especially under high-speed or high-load conditions. It is difficult to accurately detect slight changes in roll gap deviation, resulting in delayed response when deviations occur, affecting production quality and equipment maintenance cycle.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art known to those skilled in the art. Summary of the invention

[0005] The present invention provides a method, system and equipment for detecting deviation of a rolling mill roll gap, which can effectively solve the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is: A method for detecting deviation of a rolling mill roll gap, the method comprising: A vibration detection center line is established based on the copper material to be processed, and the vibration detection center line is parallel to the roll axis of the working roll; Acquire the roller vibration characteristics based on the vibration detection center line with each working roller as a unit, and generate a vibration characteristic matrix corresponding to the processing time axis of the working roller; A reference vibration deviation model is established, the vibration characteristic matrix corresponding to each working roll is subjected to vibration deviation comparison, and the roll gap deviation detection result is obtained according to the vibration deviation result mapping.

[0007] Furthermore, the vibration characteristic matrix corresponding to each of the working rolls is subjected to vibration deviation comparison, including: A plurality of vibration monitoring points are respectively set at corresponding positions of the vibration detection center line and the working roll, and the roller vibration characteristics of the working roll at the corresponding positions of the vibration detection center line are respectively obtained; Determine the duration of the processing time, and divide the duration into time slices of equal length; According to the vibration monitoring points and time slice lengths, vibration deviation comparison is performed on the roll vibration characteristics of each of the working rolls with the same coordinates and the same time slice to obtain a number of deviation comparison parameters; The vibration deviation result is obtained according to a number of the deviation comparison parameters.

[0008] Furthermore, the roller vibration characteristics are collected and acquired, including: Capturing roller vibration signals, evaluating the correlation of the roller vibration signals, and obtaining roller vibration characteristics; The roller vibration signal is collected in real time, the vibration characteristic matrix is ​​updated, the reference vibration deviation model is loaded, and the roller gap deviation detection result is output.

[0009] Furthermore, the correlation evaluation of the roller vibration signal includes: Preprocessing the roller vibration signal, performing signal decomposition on the preprocessed roller vibration signal, and obtaining a plurality of intrinsic mode functions, each of which represents a specific frequency band of the roller vibration signal; Screening a plurality of the intrinsic mode functions to obtain a plurality of relevant frequency bands, wherein the relevant frequency bands represent the intrinsic mode functions associated with the roll gap deviation; The roller vibration signal is spliced ​​and reconstructed according to the several related frequency bands to obtain the roller vibration characteristics, and the splicing and reconstruction means splicing and reorganizing the several related frequency bands.

[0010] Further, the pre-processed roller vibration signal is subjected to signal splitting, including: Pre-decomposing the roller vibration signal based on frequency to obtain a number of initial modal functions; Performing time-frequency analysis on a number of the initial modal functions, estimating frequency components of the initial modal functions, and setting an initial center frequency according to the frequency components; Collect historical rolling mill operating condition information, set a minimization objective function according to the historical rolling mill operating condition signal, iteratively optimize the frequency range of the initial modal function according to the minimization objective function, constrain the center frequency, and obtain several intrinsic modal functions, wherein the intrinsic modal functions represent different frequency components in the roll vibration signal.

[0011] Furthermore, a reference vibration deviation model is established, and the roll gap deviation detection result is obtained according to the vibration deviation result mapping, including: Constructing a roll vibration database, wherein the roll vibration database includes historical roll vibration characteristic information and historical roll gap deviation information; The roll vibration database classifies the historical roll vibration characteristic information and the historical roll gap deviation information according to the processing equipment and the copper material to be processed; Performing deep learning on the roll vibration database, and obtaining a historical mapping relationship between the historical roll vibration characteristic information and the historical roll gap deviation information of the same equipment to be processed and the copper material to be processed; The vibration deviation result is matched with the roller vibration database, and the roller gap deviation detection result is obtained according to the historical mapping relationship.

[0012] Furthermore, a roll vibration database is constructed, including: Collecting the historical roll vibration characteristic information and the historical roll gap deviation information; Clustering and managing the historical roll vibration characteristic information according to the monitoring position of the vibration detection centerline and the processing time axis, and each data item of the vibration characteristic matrix corresponds to the historical roll gap deviation information; The monitoring position, the processing time axis and the historical roll gap deviation information are respectively used as database indexes for the reference vibration deviation model to call the data link.

[0013] Further, the vibration deviation result is obtained according to a number of deviation comparison parameters, including: Setting a deviation warning value, traversing the plurality of deviation control parameters according to the deviation warning value, and obtaining a plurality of vibration parameter deviations; The time slices and position coordinates of the vibration monitoring points corresponding to the plurality of vibration parameter deviations are obtained, and the vibration deviation results are obtained.

[0014] A rolling mill roll gap deviation detection system, the system comprising: A reference vibration centerline construction module is used to establish a vibration detection centerline based on the copper material to be processed, and the vibration detection centerline is parallel to the roll axis of the working roll; A working vibration collection module collects and obtains roller vibration characteristics based on the vibration detection center line and takes each working roller as a unit, and generates a vibration characteristic matrix corresponding to the processing time axis of the working roller; The vibration deviation model comparison module establishes a reference vibration deviation model, performs vibration deviation comparison on the vibration characteristic matrix corresponding to each working roll, and obtains the roll gap deviation detection result according to the vibration deviation result mapping.

[0015] A rolling mill roll gap deviation detection device, the device applies any of the rolling mill roll gap deviation detection methods described above.

[0016] The technical solution of the present invention can achieve the following technical effects: It effectively solves the problem that it is difficult to capture slight vibration changes in roll gap deviation under high-speed and high-load conditions. Through dynamic monitoring and analysis based on vibration signals, it can capture slight vibration changes under high-speed and high-load conditions, effectively improving the sensitivity and detection accuracy of roll gap deviation. Combined with the deep learning model for data comparison, it can timely detect roll gap deviation problems, avoid the problem of delayed response in traditional methods, and significantly improve the quality control capability in the production process.

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a schematic flow chart of a method for detecting deviation of a rolling mill roll gap; Figure 2 It is a flow chart of vibration deviation control; Figure 3 Schematic diagram of the process of roller vibration feature acquisition; Figure 4 It is the relationship diagram for evaluating the correlation of the roll vibration signal; Figure 5 Split the relationship graph for the signal. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0022] Embodiment 1; like Figure 1 As shown, the present application provides a method for detecting deviation of a roll gap of a rolling mill, the method comprising: S100: establishing a vibration detection center line based on the copper material to be processed, and the vibration detection center line is parallel to the roll axis of the working roll; S200: collecting and acquiring roller vibration characteristics based on each working roller as a unit according to the vibration detection center line, and generating a vibration characteristic matrix corresponding to the processing time axis of the working roller; S300: Establish a reference vibration deviation model, compare the vibration deviation of the vibration characteristic matrix corresponding to each working roll, and obtain the roll gap deviation detection result according to the vibration deviation result mapping.

[0023] Specifically, firstly, a vibration detection center line is established as a reference. The establishment of the vibration detection center line is usually determined in combination with the processing characteristics of the copper material and the type of the working roll of the rolling mill. Specifically, during the implementation process, the precise position of the detection center line can be determined by measuring or calculating the position of the roller axis and combining the geometric characteristics of the copper material to be processed. In some embodiments, the selection of the vibration detection center line should ensure that it is parallel to the working roll axis and consistent with the processing path of the copper material; after the vibration detection center line is determined, the vibration characteristics are determined according to the vibration detection center line, and all vibration characteristics are combined into a matrix in time series, each column represents the vibration characteristics at a specific time, and each row represents The vibration characteristics at all vibration monitoring points at that moment, in order to avoid the data dimension being too large or unbalanced, usually the generated feature matrix is ​​standardized. The standardization methods may include normalization, z-score standardization, etc., so that each eigenvalue in the matrix can be compared at the same scale. In addition, the feature matrix can be reduced in dimension (such as principal component analysis) as needed to reduce redundant information and improve the efficiency of subsequent calculations. After the feature matrix is ​​generated, it will serve as the basis for deviation detection and be compared with the benchmark vibration deviation model. By comparing the vibration characteristics in the matrix with the ideal vibration characteristics in the standard model, the deviation value can be calculated, thereby realizing the detection of roller gap deviation.

[0024] The technical solution of the present invention effectively solves the problem that it is difficult to capture slight vibration changes in roll gap deviation under high-speed and high-load conditions. Through dynamic monitoring and analysis based on vibration signals, slight vibration changes under high-speed and high-load conditions can be captured, which effectively improves the sensitivity and detection accuracy of roll gap deviation. In combination with the deep learning model for data comparison, the deviation problem of the roll gap can be discovered in time, avoiding the problem of delayed response in traditional methods and significantly improving the quality control capability in the production process.

[0025] Further, if Figure 2 As shown, the vibration characteristic matrix corresponding to each working roll is subjected to vibration deviation comparison, including: S310: setting a plurality of vibration monitoring points at corresponding positions of the vibration detection center line and the working roll, and obtaining the roll vibration characteristics of the working roll at the corresponding positions of the vibration detection center line; S320: Determine the length of the processing time, and divide the time length into time slices of equal length; S330: performing vibration deviation comparison on the roll vibration characteristics of each working roll with the same coordinate and the same time slice according to the vibration monitoring point and the time slice length, and obtaining a plurality of deviation comparison parameters; S340: Obtain vibration deviation results according to a number of deviation comparison parameters.

[0026] As a preferred embodiment of the above, during the implementation process, first, vibration monitoring points are set according to the vibration detection center line. The setting of the vibration monitoring points depends on the working state of the roller and the change law of the vibration characteristics. Usually, multiple vibration sensors are arranged at key positions of each working roller (such as contact surfaces and support points). These sensors can be accelerometers, displacement sensors or piezoelectric sensors, etc., which are used to obtain the vibration characteristics of the roller in real time. Each vibration monitoring point will collect the vibration signal of the roller at a specific position. In order to fully capture the vibration changes of the roller, different monitoring point densities can be selected according to different working conditions of the rolling mill. The number and position of the monitoring points should be adjusted according to actual process requirements. For example, high-frequency sampling is used to capture high-frequency vibrations, and low-frequency sampling is used to capture low-frequency vibrations. Next, according to the processing time of the rolling mill, the time length is divided into equal-length time slices. The length of the processing time can be set according to the production cycle of the rolling mill, and the length of the time slice should be consistent with the production process. The length of each time slice can be adapted to the rhythm of the process. Generally, the length of each time slice can be set to range from a few seconds to tens of seconds. The specific length needs to be determined according to the running speed of the rolling mill, the process requirements and the characteristics of the signal change. By dividing the time length into time slices of equal length, it can ensure that the vibration characteristics in each time slice can fully reflect the working status of the working roll in this time period, and the vibration signal in each time slice is analyzed in detail. When the vibration characteristic matrix of each time slice is generated, the next step is to perform vibration deviation comparison. During the comparison process, the data in the vibration characteristic matrix is ​​first compared with the historical data model to calculate the deviation value. The deviation value of each monitoring point will reflect the difference between the vibration state of the rolling mill at that position and the ideal state. Through the above-mentioned vibration deviation comparison process, the deviation comparison parameters obtained will be used to calculate the final vibration deviation results, and these deviation results will be directly used for deviation detection and adjustment of the mill roll gap.

[0027] Furthermore, if Figure 3 As shown, the roller vibration characteristics are collected and acquired, including: S311: Capturing roller vibration signals, evaluating the correlation of the roller vibration signals, and obtaining roller vibration characteristics; S312: Collect roller vibration signals in real time, update the vibration feature matrix, load the reference vibration deviation model, and output the roller gap deviation detection result.

[0028] In this embodiment, firstly, in order to ensure high-precision capture of the vibration characteristics of the roll, this embodiment preferably uses a high-precision multi-axis accelerometer or displacement sensor. These sensors can simultaneously collect the vibration information of the roll in multiple directions, especially can capture subtle vibration changes. At the same time, in order to comprehensively collect the vibration information of the roll, the sensor is usually arranged at the key positions of the roll, such as contact points, support points and other areas. The sensor can also be arranged on both sides of the roll or at multiple angles to capture multi-dimensional vibration signals. When the rolling mill is in normal working condition, the vibration sensor will continuously and real-time collect the vibration signals of the roll. These signals include low-frequency and high-frequency vibration components generated by the roll during the processing. The frequency setting of the signal acquisition process should be based on the rolling mill. The working characteristics of the machine are adjusted according to the slight changes in the roll gap to ensure that all the vibration characteristics of the roll are covered; then the vibration signals are evaluated for correlation based on the collected vibration signals to obtain the vibration characteristics of the roll, and the vibration signals collected in real time are organized according to the time axis, and the vibration characteristic matrix is ​​updated in real time. Whenever a new vibration signal is collected, the system will update the characteristic matrix according to the time axis. With the collection and processing of signal data in each time slice, the matrix will gradually accumulate and update to reflect the real-time vibration status of the current roll. After completing the update of the vibration characteristic matrix, the system will load the benchmark vibration deviation model. The benchmark model is based on the normal roll vibration characteristics under different working conditions and provides a standard reference for comparison with real-time data.

[0029] Further, if Figure 4 As shown, the roll vibration signal is evaluated for correlation, including: Preprocessing the roller vibration signal, splitting the preprocessed roller vibration signal, and obtaining a number of intrinsic mode functions, each of which represents a specific frequency band of the roller vibration signal; Screening a number of intrinsic mode functions to obtain a number of relevant frequency bands, wherein the relevant frequency bands represent the intrinsic mode functions associated with the roll gap deviation; The roller vibration signal is spliced ​​and reconstructed according to several related frequency bands to obtain the roller vibration characteristics. The splicing and reconstruction means splicing and reorganizing several related frequency bands.

[0030] Specifically, the vibration signal of the roller collected in real time is first denoised and filtered. In order to remove high-frequency noise and low-frequency drift, a bandpass filter is used for signal preprocessing. The bandwidth of the filter needs to be selected according to the frequency range of the roller vibration signal. Generally speaking, the bandpass filter will select a frequency range to cover the main frequency band of the roller vibration and remove irrelevant noise frequency bands. Through standardization (such as Z-score standardization) or normalization (such as mapping the vibration amplitude to between 0 and 1), the data from different signal sources have the same dimension, which is convenient for subsequent analysis and comparison. Then, the preprocessed roller vibration signal is split to obtain the intrinsic mode function. By performing spectral analysis on each intrinsic mode function, the vibration amplitude and frequency components of each frequency band can be obtained. The following spectral analysis method can be used: Use fast Fourier transform to perform spectral analysis on each intrinsic mode function to obtain a frequency-amplitude spectrum. The purpose of spectral analysis is to quantify the energy distribution of the signal in each frequency range, paying special attention to the parts where the energy is concentrated or the frequency band changes significantly. In the figure, the frequency bands with more concentrated energy are often related to the main source of vibration. Therefore, when screening, those frequency bands with higher energy peaks can be selected as the key analysis objects; then, by collecting data from multiple rolling mill operations, including normal working conditions and working conditions with roll gap deviation, the change rules of each frequency band under different working conditions are determined through comparative analysis, and statistical methods such as correlation coefficient or regression analysis are used to evaluate the correlation between different frequency bands and roll gap deviation, and those parts with significant correlation between frequency band changes and roll gap deviation are screened out. Usually, the low-frequency part represents large-scale mechanical changes, while the medium and high-frequency parts may be related to small roll gap deviations; then, the intrinsic mode functions of the screened related frequency bands are spliced ​​to obtain the roller vibration characteristics. The process of splicing and reconstruction is to combine the intrinsic mode functions of the selected frequency bands. During the splicing process, the intrinsic mode functions may need to be frequency shifted, the amplitude adjusted, and other operations to ensure the time domain continuity and frequency domain consistency of the spliced ​​vibration characteristics. The reconstructed vibration characteristics usually appear as a smoother and more recognizable signal.

[0031] Furthermore, if Figure 5 As shown, the pre-processed roller vibration signal is split into signal segments, including: Pre-decompose the roller vibration signal based on the frequency to obtain several initial modal functions; Perform time-frequency analysis on several initial modal functions, estimate the frequency components of the initial modal functions, and set the initial center frequency according to the frequency components; The historical rolling mill operating condition information is collected, and the minimization objective function is set according to the historical rolling mill operating condition signal. The frequency range of the initial modal function is iteratively optimized according to the minimization objective function, and the center frequency is constrained to obtain several intrinsic mode functions. The intrinsic mode function represents the different frequency components in the roll vibration signal.

[0032] As a preferred embodiment of the above, the pre-processed roller vibration signal is first subjected to frequency pre-decomposition. The pre-decomposition process can decompose the signal into several frequency components by wavelet transform, empirical mode decomposition or Fourier transform. Each frequency component represents different frequency band information of the signal. These frequency bands cover various characteristics of roller vibration, including low-frequency mechanical vibration and high-frequency tiny roller gap deviation information. Through these methods, the original vibration signal will be disassembled into several initial mode functions, each of which represents a frequency band information of the signal. The roller vibration signal is effectively separated within a certain frequency range, and then The time-frequency analysis of the initial modal function obtained by decomposition can be carried out in the following way: the time-frequency analysis method is applied to the decomposed modal function, and the commonly used methods are short-time Fourier transform or wavelet transform, etc. The signal is divided into small time windows. The signal in each window can be assumed to be stable, which can ensure that high resolution can be obtained in both time domain and frequency domain. Subsequently, the signal in each time window is subjected to Fourier transform or wavelet transform, and its spectrum is calculated. This can obtain the frequency component corresponding to each time point, reflecting the frequency distribution of the signal at different time points. By analyzing the spectrum, the frequency components contained in the modal function can be estimated. The main frequency components contained in the modal function are shown in the spectrum of each modal function, and the energy distribution in different frequency ranges is displayed. Finally, the time-frequency analysis results are displayed as a time-frequency diagram, with the horizontal axis being time and the vertical axis being frequency. After the time-frequency analysis, the frequency components of the initial modal function are estimated based on the frequency distribution in the spectrum diagram. The frequency components reflect the main energy positions of the roller vibration signal in each frequency band. By observing the peak position of the spectrum in the time-frequency diagram, the frequency band with the most concentrated signal energy can be determined. The frequency point corresponding to this frequency band is the center frequency. Then, by collecting historical operating information, a minimization objective function is set. The role of the objective function is to obtain the center frequency through Adjust the frequency range and center frequency to find the best frequency window so that the modal function can accurately reflect the vibration characteristics of the roll gap deviation. Perform iterative optimization based on the result of minimizing the objective function. By adjusting the frequency range and center frequency of the initial modal function, gradually approach the optimal frequency band to ensure that the frequency range can accurately reflect the characteristics of the roll gap deviation. During each optimization process, the frequency components will be re-estimated and fine-tuned based on historical operating conditions to ultimately obtain the optimized intrinsic mode function. After the iterative optimization is completed, the resulting intrinsic mode function can accurately represent the different frequency components in the roll vibration signal.

[0033] Furthermore, a reference vibration deviation model is established, and the roll gap deviation detection result is obtained according to the vibration deviation result mapping, including: Constructing a roll vibration database, which includes historical roll vibration characteristic information and historical roll gap deviation information; The roll vibration database classifies historical roll vibration characteristic information and historical roll gap deviation information according to the processing equipment and the copper material to be processed; Conduct deep learning on the roll vibration database and obtain the historical mapping relationship between the historical roll vibration characteristic information and the historical roll gap deviation information of the same processing equipment and copper material to be processed; The vibration deviation results are matched with the roll vibration database, and the roll gap deviation detection results are obtained based on the historical mapping relationship.

[0034] In this embodiment, firstly, historical vibration characteristic information and historical roll gap deviation information of the rolling mill are collected and integrated to construct a roll vibration database. The historical roll vibration characteristic information includes vibration signal data, vibration amplitude, frequency component, etc. of the rolling mill under different working conditions. The roll vibration database is classified according to processing equipment and copper materials to be processed. Different equipment (such as rolling mills of different models, different types of working rolls) and different copper materials (such as copper materials with different thicknesses or surface treatments) will lead to differences in vibration characteristics. Therefore, cluster analysis (such as K-means clustering algorithm) or multidimensional data analysis method is used to classify the roll vibration characteristic information and roll gap deviation information. Each type of data corresponds to a specific working condition combination ( Such as equipment type and copper material characteristics), so that the subsequent training of the deep learning model can be optimized for different working conditions. The roller vibration database is trained by using a neural network (such as a convolutional neural network or a long short-term memory network). The training data includes historical roller vibration characteristics and corresponding roller gap deviation information. Through training, the deep learning model can learn the mapping relationship between the roller vibration characteristics and the roller gap deviation, and extract the relationship between the roller vibration characteristics and the roller gap deviation. Subsequently, the vibration deviation result is calculated through the real-time collected vibration signal, and the real-time acquired vibration deviation result is input into the trained deep learning model. The model will output the corresponding roller gap deviation detection result according to the historical mapping relationship.

[0035] Furthermore, the roll vibration database is constructed, including: Collect historical roll vibration characteristic information and historical roll gap deviation information; The historical roll vibration feature information is clustered and managed according to the monitoring position of the vibration detection centerline and the processing time axis, and the data items of each vibration feature matrix correspond to the historical roll gap deviation information; The monitoring position, processing time axis and historical roll gap deviation information are used as database indexes for the benchmark vibration deviation model to call the data link.

[0036] Specifically, the vibration of the roll is monitored by high-precision sensors (such as accelerometers and displacement sensors) to collect vibration characteristic information under different production conditions. These characteristics include vibration amplitude, frequency, phase, etc., which represent the overall vibration state of the roll, and the roll gap deviation is usually obtained through special deviation detection equipment or directly from the control system of the rolling mill. This information records the change in roll gap width, the deviation size and its change over time. The vibration characteristic information and roll gap deviation information are divided according to the time axis based on the vibration detection center line and the monitoring position. The vibration characteristics and deviation information in each time period will be classified into the category corresponding to the time period, and the vibration characteristic data will be clustered according to factors such as processing equipment and copper material type. By using K-means clustering or other clustering algorithms, the data items of each vibration characteristic matrix are classified with the corresponding historical roll gap deviation information, which is convenient for subsequent data analysis and model training. Subsequently, by assigning a unique identifier to each monitoring point, the monitoring position, processing timeline, and roll gap deviation information are used as an index field respectively, which is convenient for quickly finding and processing vibration data at a specific location in subsequent analysis.

[0037] Furthermore, the vibration deviation result is obtained according to several deviation comparison parameters, including: Setting a deviation warning value, traversing a number of deviation control parameters according to the deviation warning value, and obtaining a number of vibration parameter deviations; The time slices corresponding to the deviations of several vibration parameters and the position coordinates of the vibration monitoring points are obtained, and the vibration deviation results are obtained.

[0038] As a preferred embodiment of the above, first, it is necessary to set a deviation warning value based on historical data analysis, production conditions and experience. The deviation warning value is a threshold used to determine whether there is an abnormality in the vibration signal. The deviation warning value can be a fixed value or can be dynamically adjusted according to production conditions and different working conditions. For example, a deviation warning value can be set when the vibration amplitude of each frequency band changes by more than 5%. For specific production states (such as high-load rolling or high-speed operation), the warning value can be increased accordingly to avoid frequent triggering of unnecessary alarms; then the deviation comparison parameter is obtained to compare the deviation warning value to check whether it exceeds the deviation warning value. If a data item exceeds the warning value, the vibration deviation of the item is recorded. Each vibration parameter deviation corresponds to a specific time slice and monitoring point. When performing deviation comparison, the system will record the timestamp when the deviation occurs and the position coordinates of the vibration monitoring point. The timestamp can be used to determine the specific time when the deviation occurs, and the position coordinates can be used to determine the location where the deviation occurs; during the deviation comparison process, if some vibration parameter deviations exceed the warning value, the system will calculate the vibration deviation results for each monitoring point and time slice.

[0039] Embodiment 2: Based on the same inventive concept as the method for detecting the deviation of a rolling mill roll gap in the aforementioned embodiment, the present invention further provides a system for detecting the deviation of a rolling mill roll gap, the system comprising: The reference vibration centerline construction module establishes a vibration detection centerline based on the copper material to be processed, and the vibration detection centerline is parallel to the roll axis of the working roll; The working vibration collection module collects and obtains the roller vibration characteristics based on the vibration detection center line and takes each working roller as a unit, and generates a vibration characteristic matrix corresponding to the processing time axis of the working roller; The vibration deviation model comparison module establishes a reference vibration deviation model, compares the vibration deviation of the vibration characteristic matrix corresponding to each working roll, and obtains the roll gap deviation detection result based on the vibration deviation result mapping.

[0040] The above-mentioned adjustment system in the present invention can effectively realize a method for detecting deviation of the roll gap of a rolling mill, and the technical effects that can be achieved are as described in the above-mentioned embodiments, which will not be repeated here.

[0041] Embodiment three; Based on the same inventive concept as the method for detecting the deviation of a rolling mill roll gap in the aforementioned embodiment, the present invention also provides a device for detecting the deviation of a rolling mill roll gap, and the device applies any method for detecting the deviation of a rolling mill roll gap.

[0042] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A method for detecting deviation of a rolling mill roll gap, characterized in that: The method comprises: A vibration detection center line is established based on the copper material to be processed, and the vibration detection center line is parallel to the roll axis of the working roll; Acquire the roller vibration characteristics based on the vibration detection center line with each working roller as a unit, and generate a vibration characteristic matrix corresponding to the processing time axis of the working roller; A reference vibration deviation model is established, the vibration characteristic matrix corresponding to each working roll is subjected to vibration deviation comparison, and the roll gap deviation detection result is obtained according to the vibration deviation result mapping.

2. The method for detecting the deviation of the roll gap of a rolling mill according to claim 1, characterized in that: The vibration characteristic matrix corresponding to each of the working rolls is subjected to vibration deviation comparison, including: A plurality of vibration monitoring points are respectively set at corresponding positions of the vibration detection center line and the working roll, and the roller vibration characteristics of the working roll at the corresponding positions of the vibration detection center line are respectively obtained; Determine the duration of the processing time, and divide the duration into time slices of equal length; According to the vibration monitoring points and time slice lengths, vibration deviation comparison is performed on the roll vibration characteristics of each of the working rolls with the same coordinates and the same time slice to obtain a number of deviation comparison parameters; The vibration deviation result is obtained according to a number of the deviation comparison parameters.

3. The method for detecting the deviation of the roll gap of a rolling mill according to claim 2, characterized in that: Collect and obtain roller vibration characteristics, including: Capturing roller vibration signals, evaluating the correlation of the roller vibration signals, and obtaining roller vibration characteristics; The roller vibration signal is collected in real time, the vibration characteristic matrix is ​​updated, the reference vibration deviation model is loaded, and the roller gap deviation detection result is output.

4. The method for detecting the deviation of the roll gap of a rolling mill according to claim 3, characterized in that: The roller vibration signal is evaluated for correlation, including: Preprocessing the roller vibration signal, performing signal decomposition on the preprocessed roller vibration signal, and obtaining a plurality of intrinsic mode functions, each of which represents a specific frequency band of the roller vibration signal; Screening a number of the intrinsic mode functions to obtain a number of relevant frequency bands, wherein the relevant frequency bands represent the intrinsic mode functions associated with the roll gap deviation; The roller vibration signal is spliced ​​and reconstructed according to the several related frequency bands to obtain the roller vibration characteristics, and the splicing and reconstruction means splicing and reorganizing the several related frequency bands.

5. The method for detecting the deviation of the roll gap of a rolling mill according to claim 4, characterized in that: The pre-processed roller vibration signal is subjected to signal splitting, including: Pre-decomposing the roller vibration signal based on frequency to obtain a number of initial modal functions; Performing time-frequency analysis on a number of the initial modal functions, estimating frequency components of the initial modal functions, and setting an initial center frequency according to the frequency components; Collect historical rolling mill operating condition information, set a minimization objective function according to the historical rolling mill operating condition signal, iteratively optimize the frequency range of the initial modal function according to the minimization objective function, constrain the center frequency, and obtain several intrinsic modal functions, wherein the intrinsic modal functions represent different frequency components in the roll vibration signal.

6. The method for detecting the deviation of the roll gap of a rolling mill according to claim 1, characterized in that: Establish a benchmark vibration deviation model and obtain the roll gap deviation detection results based on the vibration deviation result mapping, including: Constructing a roll vibration database, wherein the roll vibration database includes historical roll vibration characteristic information and historical roll gap deviation information; The roll vibration database classifies the historical roll vibration characteristic information and the historical roll gap deviation information according to the processing equipment and the copper material to be processed; Performing deep learning on the roll vibration database, and obtaining a historical mapping relationship between the historical roll vibration characteristic information and the historical roll gap deviation information of the same equipment to be processed and the copper material to be processed; The vibration deviation result is matched with the roller vibration database, and the roller gap deviation detection result is obtained according to the historical mapping relationship.

7. The method for detecting the deviation of the roll gap of a rolling mill according to claim 6, characterized in that: Construct roll vibration database, including: Collecting the historical roll vibration characteristic information and the historical roll gap deviation information; Clustering and managing the historical roll vibration characteristic information according to the monitoring position of the vibration detection centerline and the processing time axis, and each data item of the vibration characteristic matrix corresponds to the historical roll gap deviation information; The monitoring position, the processing time axis and the historical roll gap deviation information are respectively used as database indexes for the reference vibration deviation model to call the data link.

8. The method for detecting the deviation of the roll gap of a rolling mill according to claim 2, characterized in that: The vibration deviation result is obtained according to a number of deviation comparison parameters, including: Setting a deviation warning value, traversing the plurality of deviation control parameters according to the deviation warning value, and obtaining a plurality of vibration parameter deviations; The time slices and position coordinates of the vibration monitoring points corresponding to the plurality of vibration parameter deviations are obtained, and the vibration deviation results are obtained.

9. A rolling mill roll gap deviation detection system, characterized in that: The system comprises: A reference vibration centerline construction module is used to establish a vibration detection centerline based on the copper material to be processed, and the vibration detection centerline is parallel to the roll axis of the working roll; A working vibration collection module collects and obtains roller vibration characteristics based on the vibration detection center line and takes each working roller as a unit, and generates a vibration characteristic matrix corresponding to the processing time axis of the working roller; The vibration deviation model comparison module establishes a reference vibration deviation model, performs vibration deviation comparison on the vibration characteristic matrix corresponding to each working roll, and obtains the roll gap deviation detection result according to the vibration deviation result mapping.

10. A rolling mill roll gap deviation detection device, characterized in that: The device applies the rolling mill roll gap deviation detection method described in any one of claims 1-8.

Images