Search

CN-121504258-B - Nonferrous metal calendaring quality management system and nonferrous metal calendaring quality management method

CN121504258BCN 121504258 BCN121504258 BCN 121504258BCN-121504258-B

Abstract

The invention relates to the technical field of metal processing quality control, in particular to a nonferrous metal calendaring quality management system and method; the system comprises a rolling working condition data acquisition unit, a hollow risk intelligent identification unit, a hollow risk classification evaluation unit, a double-branch attention time sequence neural network model, an interference fringe closed loop regulation and control unit and a rolling quality classification evaluation unit, wherein the rolling working condition data acquisition unit is used for acquiring rolling process data and microstructure data to generate rolling structure data tensor, the hollow risk intelligent identification unit is used for constructing a hollow nuclear potential image and a disturbance response image, inputting the hollow nuclear potential image and the disturbance response image into the double-branch attention time sequence neural network model, outputting the hollow nuclear potential image and the disturbance response image to obtain a hollow risk scoring matrix, the interference fringe closed loop regulation and control unit is used for constructing a pseudo film evolution image, generating a spray array response coefficient matrix and outputting the spray cooling valve group to serve as a control quantity, and the rolling quality classification evaluation unit is used for constructing a multi-grade rolling quality label set to generate a rolling quality evaluation table. The invention realizes intelligent identification of cavity risk and closed-loop quality control of cooling film thickness in the metal rolling process.

Inventors

  • Luan Xiangjie

Assignees

  • 商南县银丰铝业有限公司

Dates

Publication Date
20260508
Application Date
20251112

Claims (10)

  1. 1. A nonferrous metal calendaring quality management system, comprising: The rolling working condition data acquisition unit (1), the rolling working condition data acquisition unit (1) is used for acquiring rolling process data and microstructure data in the hot rolling and cold rolling processes, and generating rolling structure data tensors according to a time sequence of data acquisition and the spatial positions of the metal strip in the rolling direction and the transverse direction; The cavity risk intelligent recognition unit (2) is used for extracting a grain structure tensor and a tension temperature disturbance channel in the calendaring structure data tensor to construct a cavity nuclear potential diagram and a disturbance response diagram, inputting the cavity nuclear potential diagram and the disturbance response diagram into a double-branch attention time sequence neural network model, and outputting to obtain a cavity risk heat diagram and a cavity risk scoring matrix; The interference fringe closed-loop regulation and control unit (3), the interference fringe closed-loop regulation and control unit (3) is used for collecting interference image sequences of cold pressing outlets, extracting interference fringe phase transfer characteristics, reconstructing a film thickness response graph, constructing a pseudo film evolution graph based on a cavity risk heat graph, generating a spray array response coefficient matrix by calculating a film thickness phase residual function, and outputting the spray array response coefficient matrix to a spray cooling valve bank as a control quantity; The rolling quality grading evaluation unit (4) is used for constructing a multi-grade rolling quality label set by using a quality grade mapping rule based on the cavity risk scoring matrix and the spray array response coefficient matrix, and generating a rolling quality evaluation table by combining rolling structure data tensors.
  2. 2. The nonferrous metal rolling quality management system according to claim 1, wherein in the rolling working condition data acquisition unit (1), rolling process data comprise rolling force, rolling speed, rolling temperature, rolling gap and metal strip tension; the rolling structure data tensor is obtained by mapping and fusing rolling process data and microstructure data, and takes time, longitudinal position and transverse position as three-dimensional index dimensions.
  3. 3. The nonferrous metal rolling quality management system according to claim 2, wherein the void risk intelligent recognition unit (2) comprises a structural disturbance channel extraction module (21); The structure disturbance channel extraction module (21) is used for carrying out multi-channel extraction on the rolled structure data tensor to obtain a crystal grain structure tensor and a tension temperature disturbance channel; the tension temperature disturbance channel is a disturbance sequence matrix formed by tension data and temperature data in the rolling process and is constructed based on time sequence gradient change and used for representing the response of tension and temperature to a metal microstructure.
  4. 4. A nonferrous metal calendaring quality management system according to claim 3, characterized in that said void risk intelligent recognition unit (2) further comprises a void potential map construction module (22); The cavity potential mapping construction module (22) is used for constructing a cavity nuclear potential diagram according to the tensor of the grain structure and constructing a disturbance response diagram based on the tension temperature disturbance channel, and the specific method comprises the following steps: S22.1, calculating a local difference rate of grain size, a variation degree of twin crystal density and a phase boundary distribution gradient based on grain size distribution, twin crystal density and the number of phase boundaries in a grain structure tensor, and constructing a cavity nuclear potential diagram in a weighted linear combination mode; s22.2, respectively performing first-order and second-order time difference operation on the tension data and the temperature data in the tension temperature disturbance channel, and introducing a weighted response function based on the calendering transverse space coordinates to generate a disturbance response graph.
  5. 5. The nonferrous metal rolling quality management system according to claim 4, wherein the void risk intelligent recognition unit (2) further comprises a risk assessment neural network module (23); The risk assessment neural network module (23) is based on a double-branch attention time sequence neural network model, and a cavity risk heat map and a cavity risk scoring matrix are obtained by combining a cavity nuclear potential map and a disturbance response map as inputs; The double-branch attention time sequence neural network model comprises a structural feature branch and a disturbance response branch, wherein the structural feature branch is used for receiving a hollow nuclear potential image and extracting microstructure space evolution features through a graph embedding method and a space attention mechanism; the cavity risk heat map is a space distribution map and is used for indicating a cavity forming position, and the cavity risk scoring matrix is a two-dimensional risk numerical matrix of the metal strip along the length direction and the width direction and is used for quantitatively calculating cavity risks.
  6. 6. The nonferrous metal rolling quality management system according to claim 5, wherein the interference fringe closed-loop regulation unit (3) comprises an interference feature extraction module (31); The interference feature extraction module (31) is used for reconstructing a film thickness response chart, and the specific method comprises the following steps: s31.1, collecting an interference image sequence of the rolled metal strip at the outlet of the cold press, and registering and correcting distortion of the interference image sequence according to a time axis; S31.2, extracting phase transfer characteristics of interference fringes by adopting a phase shift interferometry for all images of the interference image sequence to generate a pixel-level phase distribution map; s31.3, calculating a phase gradient and a phase mutation area based on the phase distribution diagram, and generating a film thickness response diagram; The film thickness response graph is a two-dimensional space distribution graph obtained by converting fringe phase transfer characteristics in an interference image sequence through a phase thickness mapping relation and is used for representing the relative thickness distribution of a surface film layer of the metal strip at a cold pressing outlet and the local gradient change of the surface film layer; the phase shift interferometry comprises the steps of calculating interference phase distribution based on a phase shift interferometry principle, and constructing a stable fringe evolution model through multi-frame comparison to represent the spatial non-uniformity of the thickness of the surface film layers of the metal strip before and after cooling control.
  7. 7. The nonferrous metal calendaring quality management system according to claim 6, wherein the interference fringe closed-loop control unit (3) further comprises a film evolution modeling module (32); The film evolution modeling module (32) is used for constructing a pseudo film evolution graph based on the cavity risk heat map and the film thickness response map, and the specific method comprises the following steps: S32.1, taking the cavity risk heat map as a region weight map, and superposing the region weight map in a film thickness response map to construct a pseudo film evolution map; S32.2, calculating a phase error field between the film thickness response diagram and the pseudo film evolution diagram, and extracting a film thickness phase residual function; S32.3, weighting the transverse position of the metal strip through curvature normalization transformation, and adjusting the local fluctuation range of a film thickness phase residual function; The pseudo film evolution graph is a two-dimensional mapping image constructed based on a cavity risk heat map and is used for evolving a space-time variation trend of potential film thickness abnormality of a specific area in the cold pressing process; The film thickness phase residual function is a function obtained by calculating the phase difference between the film thickness response graph and the pseudo film evolution graph and is used for quantifying the deviation degree between the film thickness response and the potential abnormal region.
  8. 8. The non-ferrous metal calendaring quality management system of claim 7, wherein the interference fringe closed loop control unit (3) further includes a spray response control module (33); The spray response control module (33) is used for calculating a spray array response coefficient matrix according to a film thickness phase residual function, converting the spray array response coefficient matrix into a driving signal, and dynamically regulating and controlling the cooling flow and spray distribution of the spray cooling valve group, and the specific method comprises the following steps: S33.1, calculating a heat flux adjustment value of a corresponding area of each transverse nozzle of the metal strip according to a film thickness phase residual function; S33.2, introducing a weight function based on nozzle spacing and coverage area to construct a spray array response coefficient matrix; s33.3, outputting the spray array response coefficient matrix to a spray cooling valve group in real time, and controlling cooling flow and spray distribution.
  9. 9. The nonferrous metal rolling quality management system according to claim 8, wherein the rolling quality grading evaluation unit (4) calculates rolling quality grade grading vectors based on a void risk grading matrix and a spray array response coefficient matrix, and converts the rolling quality grade grading vectors into a rolling quality label set using a quality grade mapping rule, and generates a rolling quality evaluation table in combination with rolling structure data tensors; The rolling quality grade grading vector is based on quantitative representation of quality evaluation results of each rolling section of defect risk and regulation response, wherein the rolling quality label set is used for mapping continuous grading vectors into discrete labels and dividing the discrete labels into four grades of excellent, good, acceptable and abnormal; wherein, the rolling quality evaluation table is used for showing the quality condition of the nonferrous metal strip at each stage of rolling.
  10. 10. A nonferrous metal rolling quality management method for the nonferrous metal rolling quality management system according to any one of claims 1 to 9, characterized by comprising the steps of: S10.1, acquiring rolling process data and microstructure data in the hot rolling and cold rolling processes, and generating rolling structure data tensors according to a time sequence acquired by the data and the spatial positions of the metal strip in the rolling direction and the transverse direction; S10.2, extracting a grain structure tensor and a tension temperature disturbance channel in a calendaring structure data tensor to construct a hollow nuclear potential diagram and a disturbance response diagram, inputting the hollow nuclear potential diagram and the disturbance response diagram into a double-branch attention time sequence neural network model, and outputting to obtain a hollow risk heat diagram and a hollow risk scoring matrix; S10.3, acquiring an interference image sequence of a cold pressing outlet, extracting interference fringe phase transfer characteristics, reconstructing a film thickness response graph, constructing a pseudo film evolution graph based on a cavity risk heat graph, generating a spray array response coefficient matrix by calculating a film thickness phase residual function, and outputting the spray array response coefficient matrix to a spray cooling valve bank as a control quantity; S10.4, constructing a multi-level calendaring quality label set by using a quality level mapping rule based on the cavity risk scoring matrix and the spray array response coefficient matrix, and generating a calendaring quality evaluation table by combining the calendaring structure data tensor.

Description

Nonferrous metal calendaring quality management system and nonferrous metal calendaring quality management method Technical Field The invention relates to the technical field of metal processing quality control, in particular to a nonferrous metal calendaring quality management system and method. Background In the nonferrous metal processing field, the metal strip calendaring process is an important basic link for realizing high-precision sheet manufacture, is widely applied to the manufacturing process of foil materials, medium plates and other materials of copper, aluminum, alloys thereof and the like, along with the continuous improvement of requirements of high-end fields such as aerospace, electronic packaging, new energy batteries and the like on microstructure uniformity, surface film layer integrity and calendaring process stability of the metal strip, the calendaring quality control method which relies on artificial experience and periodic spot check is gradually difficult to meet the real-time, fine and closed-loop control requirements of modern intelligent manufacturing, therefore, an intelligent calendaring quality management system needs to be constructed; the existing nonferrous metal rolling quality control technology has three key problems that firstly, quality data in the rolling process are not comprehensively obtained, particularly, a structural mapping means for parameters of a rolling process is lacked on a microstructure level, so that defect risks of voids, inclusions and film layer abnormality cannot be identified in real time in the rolling process, secondly, most of current spray cooling control is based on experience parameter adjustment or offline film thickness measurement results, dynamic correction of a regulation and control strategy is difficult according to local film layer deviation, and thirdly, rolling quality assessment often depends on manual sampling inspection or result detection and lacks quantitative grade judgment based on process data. Disclosure of Invention The invention aims to provide a nonferrous metal calendaring quality management system and method, which are used for solving three key problems in the background technology. In order to achieve the above object, the present invention provides a nonferrous metal rolling quality management system, comprising: The rolling working condition data acquisition unit is used for acquiring rolling process data and microstructure data in the hot rolling and cold rolling processes and generating rolling structure data tensors according to a time sequence acquired by the data and the spatial positions of the metal strip in the rolling direction and the transverse direction; the cavity risk intelligent recognition unit is used for extracting a grain structure sub tensor and a tension temperature disturbance channel in the calendaring structure data tensor to construct a cavity nuclear potential diagram and a disturbance response diagram, inputting the cavity nuclear potential diagram and the disturbance response diagram into the double-branch attention time sequence neural network model, and outputting to obtain a cavity risk heat diagram and a cavity risk scoring matrix; The interference fringe closed-loop regulation and control unit is used for collecting interference image sequences of cold pressing outlets, extracting interference fringe phase transfer characteristics, reconstructing a film thickness response graph, constructing a pseudo film evolution graph based on a cavity risk thermal graph, generating a spray array response coefficient matrix by calculating a film thickness phase residual function, and outputting the spray array response coefficient matrix to a spray cooling valve bank as a control quantity; the rolling quality grading evaluation unit is used for constructing a multi-grade rolling quality label set by using a quality grade mapping rule based on the cavity risk scoring matrix and the spray array response coefficient matrix, and generating a rolling quality evaluation table by combining rolling structure data tensors. Preferably, in the rolling working condition data acquisition unit, rolling process data comprise rolling force, roller speed, roller temperature, roller gap and metal strip tension, and microstructure data comprise grain size distribution, twin crystal density and phase boundary information obtained through integrated metallographic image analysis; the rolling structure data tensor is obtained by mapping and fusing rolling process data and microstructure data, and takes time, longitudinal position and transverse position as three-dimensional index dimensions. Preferably, the cavity risk intelligent recognition unit comprises a structure disturbance channel extraction module; the structure disturbance channel extraction module is used for carrying out multi-channel extraction on the rolled structure data tensor to obtain a crystal grain structure tensor and a tension temperature dist