CN-122021348-A - Digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method

CN122021348ACN 122021348 ACN122021348 ACN 122021348ACN-122021348-A

Abstract

The invention relates to the technical field of digital twinning, in particular to a multi-physical-field simulation method of a rock wool integrated plate composite production line based on digital twinning. And then performing full-order simulation by using an SST turbulence model, a medium radiation model and a component migration equation, and obtaining a full-order distribution field of the production line by combining dynamic boundary compensation mapping plate displacement. And further extracting a space basis function by utilizing intrinsic orthogonal decomposition, and constructing a low-dimensional feature space by using a discrete empirical interpolation method and Galerkin projection. And finally, projecting the measured data to a low-dimensional space by using a parameterized background data weak observer and Gaussian process regression to reconstruct and fit coefficients, so as to realize self-correcting twin field reconstruction. The invention solves the problem that the simulation precision and the real-time performance are difficult to be compatible under the complex production line working condition, and improves the prediction reliability of the simulation.

Inventors

SUN XIN
ZHANG JIANGUO
Qiao Xiaole
ZHAO YANJUAN

Assignees

南京彤天岩棉有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (7)

1. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method is characterized by comprising the following steps of: constructing a representative volume element reflecting the microstructure of the rock wool fiber by utilizing a random fiber filling algorithm and a fractal geometric theory, and establishing a parameterized structural model based on a porosity evolution rule; by means of The SST turbulence model simulates high-speed circulating air jet flow in the curing furnace, the parameter response library is called as a porous medium region variable parameter source item, a medium radiation model and porous medium component migration are integrated and participated, dynamic boundary compensation is combined, and the displacement of a plate on a production line is mapped to obtain a full-order distribution field of the production line; Extracting a space basis function of the full-order distribution field of the production line by utilizing intrinsic orthogonal decomposition, and projecting a control equation to a feature subspace by a discrete empirical interpolation method and Galerkin projection to obtain a low-dimensional feature space; And projecting sensor data to the low-dimensional feature space by using a parameterized background data weak observer and Gaussian process regression to perform coefficient reconstruction and least square fitting to obtain a self-correcting twin field.
2. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method according to claim 1, wherein the steps of constructing a representative volume element reflecting a rock wool fiber microstructure and constructing a parameterized structure model based on a porosity evolution rule by using a random fiber filling algorithm and a fractal geometric theory include the following steps: the method comprises the steps of generating an initial fiber central axis set in a set three-dimensional Euclidean space according to the average diameter and length-diameter ratio distribution of rock wool fibers, controlling anisotropic distribution of the central axis set based on a fiber orientation probability density function set by line pendulum frequency, carrying out multi-level bifurcation and surface roughness iteration on single fibers by combining fractal geometric theory to generate fiber geometric entities with interpenetrating structures to form initial representative volume elements, establishing a geometric compression transformation operator describing spatial deformation of a fiber skeleton by taking real-time pressure distribution of line compression rollers as driving variables, and carrying out space scaling mapping on the initial representative volume elements by utilizing the geometric compression transformation operator to dynamically adjust the number of contact points among the fibers and the roundabout degree of pores to obtain a parameterized structural model evolving in real time along with the line production efficiency Kuang Kongxi.
3. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method according to claim 1, wherein the step of constructing a parameter response library comprises the following steps: The method comprises the steps of discretizing a representative volume element into a three-dimensional lattice point space, respectively constructing a speed distribution function describing fluid movement and a temperature distribution function describing energy transmission, applying a pressure gradient and a temperature difference boundary in the lattice point space, carrying out iterative computation through collision and migration operators until a speed field and a temperature field in micro-pores are converged, calculating local stress and heat flux of a fluid-solid interface by utilizing a momentum exchange method, carrying out space integral averaging on the converged micro-scale field, calculating according to a linear mapping relation between an average filtering speed and the pressure gradient to obtain an anisotropic permeability tensor, calculating according to a mapping relation between heat flux density distribution and a macroscopic temperature gradient to obtain an equivalent heat conductivity tensor, obtaining a fiber-fluid heat exchange coefficient according to a proportional relation between total heat exchange capacity of the fluid-solid interface and an average temperature difference of a body, repeatedly executing the steps through changing the porosity of a parameterized structure model, and establishing a corresponding relation table taking the anisotropic permeability tensor, the equivalent heat conductivity tensor and the fiber-fluid heat exchange coefficient as dependent variables to form the parameter response library.
4. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method of claim 1, wherein the step of obtaining the production line full-order distribution field comprises the following steps: The method comprises the steps of establishing a continuous equation and a momentum equation based on an SST turbulence model to describe a wind speed field in a curing furnace, synchronously establishing an energy relation between an integrated participation medium radiation item and a component migration item to describe a heat mass exchange process inside and on the surface of a plate, defining the plate area as a porous medium area, calling corresponding values in a parameter response library in real time according to the current porosity of the parameterized structural model to serve as a resistance source item in the momentum equation and a heat conduction and heat exchange correction item in the energy equation, updating the geometric position of the plate according to the actual conveying speed of a production line by utilizing a coordinate transformation operator, compensating flow field information in the curing furnace to the surface boundary of the moving plate, then carrying out full-order value solution, outputting and polymerizing to obtain a production line full-order distribution field containing a surface heat flux map, an internal water content gradient and a topological space wind pressure gradient line.
5. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method of claim 1, wherein the step of obtaining the low-dimensional feature space comprises the steps of: Combining all-order distribution fields of a production line under a plurality of time steps into a snapshot matrix, carrying out singular value decomposition on the snapshot matrix by utilizing intrinsic orthogonal decomposition, extracting a space basis function reflecting main components of physical evolution of the production line, constructing a low-dimensional linear subspace, identifying a nonlinear term in a control equation, selecting sparse sampling points in a physical space by utilizing a discrete empirical interpolation method, constructing an approximate projection basis of a nonlinear operator according to the space basis function, mapping all-order momentum equations and energy equations to the low-dimensional linear subspace by utilizing the space basis function and the approximate projection basis of the nonlinear operator through Galerkin projection, eliminating the dependence of calculation complexity on all-order grid degrees of freedom, and obtaining the low-dimensional feature space.
6. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method of claim 1, wherein the step of obtaining the self-correcting twinning field comprises the following steps: The method comprises the steps of using a parameterized background data weak observer, taking discrete measurement point data acquired by a sensor as a constraint item, executing variation assimilation calculation in the low-dimensional feature space to extract an initial modal coefficient matched with a current actual measurement working condition, predicting a nonlinear residual error between the low-dimensional feature space and a full-order distribution field by using a Gaussian process regression model to generate a calibration operator for compensating an unmodeled physical effect of a system, linearly combining the initial modal coefficient and the calibration operator by using a least square fitting operator to reconstruct a corrected dynamic modal coefficient, projecting the corrected dynamic modal coefficient back to the space basis function, and synthesizing a self-correcting twin field with physical conservation characteristics and real-time correction.
7. The digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method of claim 1, wherein the self-rectifying twinning field is a real-time dynamic distribution field with physical conservation constraint, and the synthesis process comprises the following steps: and projecting the corrected dynamic modal coefficient back to a characteristic subspace defined by a space basis function, obtaining a grid node numerical value set under a full-order calculation domain through matrix linear reconstruction and reduction, and displaying the temperature field evolution process and the gradient distribution of the water content in the rock wool board in real time at a monitoring terminal in the form of a three-dimensional voxel cloud picture.

Description

Digital twinning-based rock wool integrated plate composite production line multi-physical field simulation method Technical Field The invention relates to the technical field of digital twinning, in particular to a rock wool integrated plate composite production line multi-physical field simulation method based on digital twinning. Background In the rock wool integrated plate production process, the traditional numerical simulation method is mainly based on static physical parameters to perform rough calculation, and it is difficult to accurately describe the dynamic evolution rule of the internal micro-pore structure of the plate in the high-temperature wind field of the curing furnace. The existing service scheduling system and simulation platform lack of real-time characterization capability for porous medium heat, flow, mass and radiation multi-field coupling process, so that obvious deviation exists between simulation results and actual production line, and real-time requirements of sub-second self-deviation correction and fine management and control in the high-speed production process are difficult to meet. The traditional simulation method has the defects that key physical fields such as heat energy transmission, component migration, medium radiation and the like are in a logic island state in numerical calculation, and high-fidelity collaborative solving of multiple physical fields is difficult to realize. And the simulation framework based on the full-order full grid has overlarge degree of freedom, when iterative computation involving turbulent flow source and radiation nonlinear residual error is processed, the computation complexity is exponentially increased, and the sub-second response requirement of a controlled object of a production line which is far more time-consuming in resolving is solved, so that real-time numerical deduction is not feasible in engineering practice. In order to solve the technical problems of dynamic deduction and self-correction with high physical reliability and meeting real-time requirements for the spatial morphological evolution and internal state of the whole life cycle of rock wool production in a complex multi-physical-field coupling environment, a rock wool integrated plate composite production line multi-physical-field simulation method based on digital twinning is provided. Disclosure of Invention The invention aims to provide a digital twinning-based rock wool integrated plate composite production line multi-physical-field simulation method, which constructs a multi-scale simulation framework of a deep fusion microcosmic physical property characterization and macroscopic reduced order model and realizes real-time accurate prediction and virtual-real dynamic correction of the state of the rock wool production line multi-physical-field. In order to achieve the above purpose, the present invention provides the following technical solutions: a rock wool integrated plate composite production line multi-physical field simulation method based on digital twinning comprises the following steps: constructing a representative volume element reflecting the microstructure of the rock wool fiber by utilizing a random fiber filling algorithm and a fractal geometric theory, and establishing a parameterized structural model based on a porosity evolution rule; Simulating high-speed circulating air jet flow in a curing furnace by using an SST turbulence model, calling the parameter response library as a porous medium region variable parameter source item, integrating a medium radiation model and porous medium component migration, and mapping the displacement of a plate on a production line by combining dynamic boundary compensation to obtain a full-order distribution field of the production line; Extracting a space basis function of the full-order distribution field of the production line by utilizing intrinsic orthogonal decomposition, and projecting a control equation to a feature subspace by a discrete empirical interpolation method and Galerkin projection to obtain a low-dimensional feature space; And projecting sensor data to the low-dimensional feature space by using a parameterized background data weak observer and Gaussian process regression to perform coefficient reconstruction and least square fitting to obtain a self-correcting twin field. Preferably, the method utilizes a random fiber filling algorithm and a fractal geometric theory to construct a representative volume element reflecting the microstructure of the rock wool fiber, and establishes a parameterized structural model based on a porosity evolution rule, and the method comprises the following steps: the method comprises the steps of generating an initial fiber central axis set in a set three-dimensional Euclidean space according to the average diameter and length-diameter ratio distribution of rock wool fibers, controlling anisotropic distribution of the central axis set based on a fiber orientation pro