Search

CN-121997184-A - Automatic control system and method for calender

CN121997184ACN 121997184 ACN121997184 ACN 121997184ACN-121997184-A

Abstract

The application relates to the field of intelligent control of calenders, and particularly discloses an automatic control system and method of a calender, which are characterized in that firstly, surface monitoring videos of rollers in the calender to be monitored, temperature values of a plurality of preset time points of the rollers in the calender to be monitored and product quality monitoring images of a plurality of preset time points of products to be monitored are collected, then, the surface monitoring videos, the temperature values of the preset time points and the product quality monitoring images of the preset time points are respectively subjected to convolution coding through a convolution neural network model in deep learning to obtain roller surface state feature vectors, roller temperature change feature vectors and product quality feature vectors, and then, the roller surface state feature vectors, the roller temperature change feature vectors and the product quality feature vectors are subjected to feature fusion to obtain classification results for indicating whether the working state of the calender to be monitored needs to be adjusted.

Inventors

  • ZHOU XING

Assignees

  • 湖州长星金属制品有限公司

Dates

Publication Date
20260508
Application Date
20240105

Claims (10)

  1. 1. An automated control system for a calender, comprising: The calender data acquisition module is used for acquiring surface monitoring videos of rollers in the calender to be monitored, temperature values of a plurality of preset time points of the rollers in the calender to be monitored and product quality monitoring images of a plurality of preset time points of a product to be monitored; The calender data preprocessing module is used for respectively preprocessing the surface monitoring video, the temperature values of the plurality of preset time points and the product quality monitoring images of the plurality of preset time points to obtain a three-dimensional roller surface state video tensor, a one-dimensional roller surface temperature input vector and a three-dimensional product quality input tensor; the calender data characteristic extraction module is used for respectively extracting characteristic information of the three-dimensional roller surface state video tensor, the one-dimensional roller surface temperature input vector and the three-dimensional product quality input tensor to obtain a roller surface state characteristic vector, a roller temperature change characteristic vector and a product quality characteristic vector; And the classification result generation module is used for obtaining a classification result based on the roller surface state characteristic vector, the roller temperature change characteristic vector and the product quality characteristic vector.
  2. 2. The automated control system of a calender of claim 1, wherein the calender data preprocessing module comprises: the video preprocessing unit is used for acquiring key information in the surface monitoring video to obtain the three-dimensional roller surface state video tensor; A temperature value preprocessing unit for arranging the temperature values of the plurality of predetermined time points into the one-dimensional roller surface temperature input vector; And the image preprocessing unit is used for arranging the product quality monitoring images at a plurality of preset time points into the three-dimensional product quality input tensor.
  3. 3. The automated control system of a calender according to claim 2, wherein the video preprocessing unit comprises: The video key frame acquisition subunit is used for acquiring a plurality of roller surface state video key frames from the surface monitoring video; and the key frame arrangement subunit is used for arranging the plurality of roller surface state video key frames into the three-dimensional roller surface state video tensor according to the time dimension.
  4. 4. The automated control system of a calender of claim 3, wherein the calender data feature extraction module comprises: The target detection unit is used for acquiring characteristic information of a roller surface state in the three-dimensional roller surface state video tensor so as to obtain a roller surface state characteristic vector; A temperature characteristic extraction unit for passing the one-dimensional roller surface temperature input vector through a roller surface temperature characteristic extraction module based on a time sequence encoder to obtain the roller temperature variation characteristic vector; And the product quality feature extraction unit is used for enabling the three-dimensional product quality input tensor to pass through a convolutional neural network model serving as a product quality feature extractor to obtain the product quality feature vector.
  5. 5. The automated control system of a calender of claim 4, wherein the target detection unit comprises: the target detection subunit is used for carrying out feature extraction on the three-dimensional roller surface state video tensor by a roller surface state feature extraction module based on a target detection network so as to obtain a roller surface state feature map; And the feature map dimension reduction subunit is used for carrying out pooling operation on the roller surface state feature map to obtain a roller surface state feature vector.
  6. 6. The automated control system of a calender of claim 5, wherein the object detection network is an anchor window based object detection network.
  7. 7. The automated control system of a calender of claim 6, wherein the timing encoder comprises a fully-connected layer and a one-dimensional convolutional layer.
  8. 8. The automated control system of a calender of claim 7, wherein the classification result generation module comprises: the roller characteristic combining unit is used for fusing the roller surface state characteristic vector and the roller temperature change characteristic vector to obtain a roller working state characteristic vector; the calender characteristic fusion unit is used for carrying out characteristic fusion on the characteristic vector of the working state of the roller and the characteristic vector of the product quality so as to obtain a classification characteristic vector of the calender; the probability unit is used for inputting the classified feature vector of the calender into a Sigmoid function to obtain a classified feature vector of the probabilistic calender; the calender characteristic optimization unit is used for carrying out rigidity consistency on parameterized geometric relationship transition priori characteristics on the classified characteristic vectors of the probabilistic calender so as to obtain optimized classified characteristic vectors of the calender; And the classification unit is used for carrying out feature classification on the classification feature vector of the optimized calender through a classifier to obtain a classification result used for indicating whether the working state of the roller in the calender to be monitored needs to be adjusted.
  9. 9. The automated control system of a calender according to claim 8, wherein the calender feature optimization unit is configured to perform a rigid equalization of parameterized geometric relationship transition prior features on the probabilistic calender classification feature vector to obtain the optimized calender classification feature vector with the following formula; Wherein, the formula is: Wherein V ij represents the eigenvalue of the (i, j) th position in the probabilistic calender classification eigenvector, λ represents a predetermined hyper-parameter, log represents a logarithmic function value based on 2, and V ij ' represents the eigenvalue of the (i, j) th position in the optimized calender classification eigenvector.
  10. 10. An automated control method of a calender, comprising: collecting surface monitoring videos of rollers in a calender to be monitored, temperature values of a plurality of preset time points of the rollers in the calender to be monitored and product quality monitoring images of a plurality of preset time points of a product to be monitored; Preprocessing the surface monitoring video, the temperature values of the plurality of preset time points and the product quality monitoring images of the plurality of preset time points respectively to obtain a three-dimensional roller surface state video tensor, a one-dimensional roller surface temperature input vector and a three-dimensional product quality input tensor; Respectively extracting characteristic information of the three-dimensional roller surface state video tensor, the one-dimensional roller surface temperature input vector and the three-dimensional product quality input tensor to obtain a roller surface state characteristic vector, a roller temperature change characteristic vector and a product quality characteristic vector; And obtaining a classification result based on the roller surface state characteristic vector, the roller temperature change characteristic vector and the product quality characteristic vector.

Description

Automatic control system and method for calender Technical Field The application relates to the field of intelligent control of calenders, and in particular relates to an automatic control system and method of a calender. Background A Rolling mill (Rolling mill) is a mechanical device used for metal working, mainly for Rolling metal materials into products such as sheets, strips, bars, etc. of different shapes and sizes. It is made by passing a metallic material through a series of rolling processes, subjecting it to plastic deformation, thus changing its cross-sectional shape and reducing its thickness. During the working process of the calender, the working state of the rollers influences the quality of the calendered product. For example, rolls are subjected to significant pressure and friction during calendering, and after prolonged use, wear may occur which may lead to inaccurate product dimensions or uneven thickness if the rolls wear unevenly, wear and deformation of the rolls may lead to increased or uneven gaps, uneven product thickness or inaccurate dimensions if the rolls gaps are incorrect, and poor product surface quality, such as burns or overheating, if the rolls temperature control is inaccurate in some particular calendering applications. Accordingly, an automated control system and method for a calender is desired that maintains product quality stable by controlling the operating conditions of the rolls (including roll wear, gap size, and surface temperature) in the calender based on deep learning techniques. Disclosure of Invention The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an automatic control system and method of a calender, which are characterized in that firstly, a surface monitoring video of a roller in the calender to be monitored, temperature values of a plurality of preset time points of the roller in the calender to be monitored and product quality monitoring images of a plurality of preset time points of a product to be monitored are collected, then, the surface monitoring video, the temperature values of the preset time points and the product quality monitoring images of the preset time points are respectively subjected to convolution coding through a convolution neural network model in deep learning to obtain a roller surface state feature vector, a roller temperature change feature vector and a product quality feature vector, and then, the roller surface state feature vector, the roller temperature change feature vector and the product quality feature vector are subjected to feature fusion to obtain a classification result for indicating whether the working state of the roller in the calender to be monitored needs to be adjusted. According to a first aspect of the present application, there is provided an automated control system for a calender, comprising: The calender data acquisition module is used for acquiring surface monitoring videos of rollers in the calender to be monitored, temperature values of a plurality of preset time points of the rollers in the calender to be monitored and product quality monitoring images of a plurality of preset time points of a product to be monitored; The calender data preprocessing module is used for respectively preprocessing the surface monitoring video, the temperature values of the plurality of preset time points and the product quality monitoring images of the plurality of preset time points to obtain a three-dimensional roller surface state video tensor, a one-dimensional roller surface temperature input vector and a three-dimensional product quality input tensor; the calender data characteristic extraction module is used for respectively extracting characteristic information of the three-dimensional roller surface state video tensor, the one-dimensional roller surface temperature input vector and the three-dimensional product quality input tensor to obtain a roller surface state characteristic vector, a roller temperature change characteristic vector and a product quality characteristic vector; And the classification result generation module is used for obtaining a classification result based on the roller surface state characteristic vector, the roller temperature change characteristic vector and the product quality characteristic vector. According to a second aspect of the present application, there is provided an automated control method of a calender, comprising: collecting surface monitoring videos of rollers in a calender to be monitored, temperature values of a plurality of preset time points of the rollers in the calender to be monitored and product quality monitoring images of a plurality of preset time points of a product to be monitored; Preprocessing the surface monitoring video, the temperature values of the plurality of preset time points and the product quality monitoring images of the plurality of preset time points respectively t