CAE (Computer Aided Engineering) design method for glass molded surface formed outside furnace

Abstract

The invention relates to the field of tools, in particular to a CAE (Computer Aided Engineering) design method for a glass molded surface formed outside a furnace, which is realized by adopting three types of software of Matlab, Python and Abaqus cooperatively, and comprises the following steps: S1, establishing an initial glass forming simulation model: performing quadrilateral mesh division (the mesh size is 5mm) on glass and a cold ring in Hypermesh, and outputting an inp format file; the inp file is imported into Abaqus, glass material attributes (including elasticity modulus, Poisson's ratio and viscoelastic parameters) are set, and the CAE design method for the glass profile formed outside the furnace checks deviation changes in real time, discovers calculation anomalies in time and avoids invalid iteration; an existing plate glass iteration method is compatible, and extra development is not needed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of tooling, in particular to a CAE design method for glass profile of furnace outside forming. BACKGROUND

[0002] In the past, the sunroof and back glass of the car generally used DB4 forming method. In recent years, driven by energy saving and emission reduction, Saint-Gobain began to produce car sunroofs and back glass in the form of furnace outside The car glass (sunroof, back glass) generally adopts DB4 press forming. DB4 press forming refers to the process of heating flat glass to about 100℃ above the glass transition temperature, pressing the convex and concave dies together, and then dropping onto the cold ring for gravity forming, and obtaining the target profile glass through air grid quenching. Among them, a plurality of air holes are arranged on the convex die, and the air holes form a semi-vacuum area to adsorb the flat glass to the die surface. The middle region of the glass is consistent with the convex die surface under the action of pressure, and the edge region of the glass cannot be completely attached to the convex die under the action of pressure. Further pressing the convex die and the die together makes the edge profile of the glass consistent with the convex die surface. Then the convex die releases the glass and forms on the cold ring under the action of gravity. Since the glass is a viscoelastic material, the forming profile of the glass on the cold ring is directly related to the convex die surface, and the convex die surface is calculated according to the theoretical glass profile.

[0003] Prior art: Fuyao's patent CN112784456A "Automobile glass forming die surface design method and system" is the prior art. The deviation of the glass profile obtained by simulation from the standard glass profile is manually corrected and compensated, and then simulated again. This cycle continues until the deviation reaches an acceptable range. That is, a first simulation glass profile is obtained by simulating the standard glass; whether the first simulation glass profile meets the design requirements is determined according to the solid model of the standard glass; in the case that the first simulation glass profile does not meet the design requirements, the first forming simulation geometric model is modified according to the solid model of the standard glass and the first simulation glass profile to obtain a second forming simulation geometric model; a second simulation glass profile is obtained by simulating the second forming simulation geometric model; whether the second simulation glass profile meets the design requirements is determined according to the solid model of the standard glass; and in the case that the second simulation glass profile meets the design requirements, the second model surface is determined as the die surface for glass forming.

[0004] Disadvantages of prior art: (1) To obtain as accurate convex die surface as possible, 8-15 iterations are required, which requires repeated manual data processing, simulation preprocessing, and simulation glass profile and standard glass profile deviation calculation, etc., which consumes a lot of time and effort, and is prone to errors.

[0005] (2) The deviation between the simulation glass profile and the standard glass profile is not visualized.

[0006] (3) The design of the theoretical glass profile is changed in the early stage of the customer, which leads to the repeated design of the convex model surface by manual work, and a large amount of time and energy is consumed.

[0007] Therefore, it is necessary to provide an off-line forming glass profile CAE design method to solve the above technical problems. SUMMARY

[0008] To solve the above technical problems, the present application provides an off-line forming glass profile CAE design method, which is realized by cooperating Matlab, Python and Abaqus software, and the specific steps are as follows: S1: Establishing an initial glass forming simulation model 1. Meshing: Import the glass and cold ring CAD models in Hypermesh, divide the meshing by quadrilateral elements (glass meshing 5mm, cold ring meshing 10mm), and output the inp file (Abaqus recognizable format) of the glass and cold ring; 2. Abaqus model setting: Material properties: glass is assigned with elastic modulus (60~70GPa), Poisson's ratio (0.2~0.25), and viscoelastic parameters (Prony series based on forming temperature); the cold ring is not set with material properties; Element type: glass is S4R deformed body shell element, and cold ring is R3D4 rigid body shell element; the center of the cold ring is set with a reference point, and all nodes are associated by "rigid connection"; Analysis step: create a new "gravity forming analysis step" (time 10~30s), apply gravity load (9.81m / s²), set forming temperature (50~100℃ above glass transition temperature), and set field output displacement, stress and temperature; 1. Output initial inp: after checking the model, output the initial iteration inp file (named "project name_S0.inp", S0 is the standard glass surface).

[0009] S2: Write Matlab calling script The core function is "call Abaqus + trigger Python", and the function runabaqus(path, InpFile, cpus) is defined: • path: program running path (such as "D:\GlassMold\Project"); • InpFile: the inp file name to be submitted (e.g., "Sunroof_S0.inp"); • cpus: the number of cores Abaqus occupies (2-8 cores); • The function calls the Abaqus command line through Matlab's system command, displays the simulation start / end time and time consumption, and facilitates the detection of calculation abnormalities.

[0010] S3: Write Python data processing scripts The core function is "read simulation results + calculate new mold surface + generate new inp": 1. Read data: open the Abaqus odb result file through the odbAccess module, extract the glass node coordinates after simulation (Sn1, the current iteration forming surface); read the standard glass surface S0 data (node number one-to-one correspondence) and the previous iteration data Sn2; 2. Iterative formula calculation: Basic formula: Sn_x = S0_x + Sn2_x - Sn1_x, Sn_y = S0_y + Sn2_y - Sn1_y, Sn_z = S0_z + Sn2_z - Sn1_z; Accelerate convergence: introduce coefficient λ (1<λ<2, preferably 1.5), the formula is updated to Sn = S0 + λ*(Sn2 -Sn1); 1. Generate new inp: replace the current inp file's glass node coordinates with Sn data, and save it as "project name_S (iter+1).inp".

[0011] S4: Write Matlab main function script As the "iteration control core", set parameters and implement loops: 1. Parameter initialization: initial iteration iter=0, maximum iteration Max_iter=20, difference threshold 0.02mm, project name, running path; 2. Loop control: through the while loop to realize the "simulation - data processing - deviation calculation - monitoring" closed loop: If iter < Max_iter and "center point difference> threshold": call runabaqus to submit simulation, call Python to process data, and calculate the center point difference (the Euclidean distance between the glass center point and the center point of S0); Real-time visualization: draw the "iteration number - center point difference" line chart to intuitively monitor the convergence progress; 1. Termination condition: stop iteration when difference is less than or equal to threshold value or iter is greater than or equal to Max_iter.

[0012] S5: Extract the mold surface data After the iteration is completed, the node coordinates of the glass in the inp file of the last iteration (such as "Sunroof_S9.inp") are extracted, which is the target mold surface data and can be directly used for mold processing.

[0013] In addition, the present application is compatible with the existing flat glass iteration method: only the initial mold surface of S1 is replaced with the flat glass flattened surface, and the iteration formula reference surface is adjusted.

[0014] Compared with the related art, the present application has the following beneficial effects: 1. Efficiency improvement: fully automated iteration, design cycle is shortened from 1-2 weeks to 1-2 days, and iteration times are reduced to 5-10 times; 2. Precision improvement: eliminate human error, mold surface deviation is controlled within 0.02mm, and glass scrap rate is reduced by more than 30%; 3. Visual monitoring: real-time view of deviation change, timely detection of calculation abnormalities, and avoidance of invalid iteration; 4. High compatibility: compatible with existing flat glass iteration method, no additional development is required; 5. Fast response to changes: only S0 data needs to be updated when the customer changes, and the main function needs to be run again, and the response time is shortened to 1 / 3 of the original. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a flowchart of the present application; Figure 2 is a specific flowchart of the present application Figure 3 is a call relationship block of matlab, python and abaqus of the present application; Figure 4 is a real-time graphical monitoring effect diagram of the difference between the center point coordinates of the molded glass surface and the theoretical glass surface center point coordinates. DETAILED DESCRIPTION

[0016] The present application will be further described below in conjunction with the drawings and embodiments.

[0017] Please refer to Figures 1 to 2 , a CAE design method for glass surface of off-line forming, which is realized by cooperation of Matlab, Python and Abaqus, and the specific steps are as follows: Step 1: In hypermesh, the glass and the cold ring are meshed, the mesh is quadrilateral, the mesh size is 5mm, and the inp format file is output, the inp file of the glass and the cold ring is imported in abaqus, the material property parameters of the glass are established, including elastic modulus, poisson's ratio, viscoelasticity parameters, etc.; the glass and the cold ring are both set as shell elements, the glass is given the material properties of the glass, the cold ring is set as a rigid body, a reference point is set on the cold ring, the reference point is connected with the cold ring by rigidity, the cold ring does not need to be given material properties, the gravity forming analysis step of the glass falling on the cold ring is established, the glass forming time is set, the glass forming temperature and the field output parameters are set, the model is checked in abaqus, the inp file is output, the inp file is copied to the folder of the iteration program, the name is changed to "project name_S(iter).inp", with the increase of the iteration iter, the calling program also changes.

[0018] Second step: A matlab script file is written, which supports running abaqus in matlab and submitting the initial iteration model inp file, and can support calling the python script file of step 3. The detailed process is: a function is written to call abaqus in matlab, which realizes the function of calling the iter_inp file of glass forming, and displays the program calculation time to detect whether the calculation process is abnormal.

[0019] For example: function [output]=runabaqus(path,InpFile,cpus) Where path is the program running path, InpFile is the "project name_S(iter).inp" file to be submitted, cpus is the number of cores occupied during running.

[0020] Third step: A python script file is written, which supports reading the result odb file of abaqus running, the initial glass state is S0, the data of S0 glass in the coordinate system is (S0_x, S0_y, S0_z), the data (S1_x, S1_y, S1_z) in the result file of S0 forming after the first iteration of S0 is extracted, and compared with the standard glass surface data and the difference is calculated to get the second iteration of the mold surface S2, here S2=S0+S0-S1, and so on, the nth iteration has Sn=S0+S(n-2)-S(n-1), the above S1 to Sn coordinate point numbers correspond one by one, so that Sn can be calculated: Sn_x = S0_x + Sn2_x - Sn1_x Sn_y = S0_y + Sn2_y - Sn1_y Sn_z = S0_z + Sn2_z - Sn1_z Until the deviation of the glass profile after simulation forming and the standard glass profile reaches the specified requirement, the coefficient 1<λ<2 can be set, Sn=S0+λ*(Sn2-Sn1), which can accelerate the convergence speed of the program. For example, the data of the glass forming result file "project name_iter.odb" is read by using the openodb module of python odbAccess, the data is processed according to the formula Sn=S0+Sn2-Sn1, and the glass data in "project name_iter.inp" is replaced and stored as "project name_(iter+1).inp". In this way, the program is called, read and stored repeatedly, so as to guide the convergence of the program.

[0021] Fourth step: A matlab main function script is written, in which a loop program is set, the iteration number and the center point difference (the center point of the glass after forming and the center point of the standard glass) are used as the convergence condition, the script in the second step is called repeatedly, the data is processed and calculated and stored, and the center point difference is plotted against the iteration number, so as to detect the convergence in real time. The path of the program running is set, the project name variable is set according to the actual project, the forming time variable is set, the initial iteration iter=0 is set, and the maximum number of iterations Max_iter is set. When iter<Max_iter, the program continues to execute the iteration process, the iteration process includes running the iter_inp file of abaqus and reading the iter_odb file name, running the calling function to call the python file to read the odb file to extract the data, and the difference between the center point coordinates of the glass profile after forming and the center point coordinates of the theoretical glass profile is monitored in real time, so as to monitor the convergence of the whole iteration process. The monitoring effect is as follows Figure 4 , after 9 iterations, the convergence is achieved, and the difference between the center point coordinates of the glass profile after forming and the center point coordinates of the theoretical glass profile is plotted in the graph, the difference between the monitoring point and the standard glass profile is monitored, and of course the monitoring point can also be changed to the maximum value of the difference between the glass profile point after forming and the theoretical glass profile point corresponding to the number.

[0022] STEP5: After the iteration is completed, the data of the inp file of the last iteration is extracted, that is, the data of the mold surface.

[0023] The above only describes the embodiments of the present application, and does not limit the patent scope of the present application, any equivalent structure or equivalent process transformation using the content of the specification and drawings, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method of off-line forming a glass profile CAE design, characterized by, Coordinated implementation by Matlab, Python and Abaqus, including the following steps: S1: Establish an initial glass forming simulation model: divide the glass and cold ring into quadrilateral meshes in Hypermesh (mesh size 5mm), output the inp format file; import the inp file into Abaqus, set the glass material properties (including elastic modulus, Poisson's ratio, viscoelastic parameters), element type (glass as a deformed shell element, cold ring as a rigid shell element) and gravity forming analysis step parameters (forming time 10~30s, forming temperature 50~100℃ above the glass transition temperature), establish the initial iteration model (initial mold surface is the standard glass surface S0), output the initial inp file; S2: Write a Matlab calling script: write a Matlab script containing the runabaqus(path, InpFile, cpus) function, which can call Abaqus to submit the inp file simulation and display the calculation time, while supporting the calling of step S3 Python script; S3: Write a Python data processing script: write a Python script to read the Abaqus simulation odb file through the odbAccess module, extract the current iteration forming glass surface data Sn1, combine the standard glass surface S0 and the previous iteration data Sn2, calculate the new mold surface Sn according to the formula Sn=S0+Sn2-Sn1 (optional introduction of 1<λ<2 coefficient, formula updated to Sn=S0+λ*(Sn2-Sn1)), generate a new inp file; S4: Write a Matlab main function script: set the initial iteration number iter=0, the maximum number of iterations Max_iter=20 and the center point difference threshold 0.02mm, call the scripts of steps S2~S3 through the while loop, real-time draw the center point difference-iteration number curve to monitor the convergence, and stop iteration when the convergence condition is met; S5: Extract the mold surface data: after iteration, extract the glass surface data in the inp file of the last iteration, which is the target mold surface data.

2. The isopipe glass profile CAE design method according to claim 1, wherein, In step S1, a reference point is set on the cold ring, which is associated with the cold ring through rigid connection constraint; the field output parameters of the Abaqus model include the displacement, stress and temperature of the glass nodes.

3. The isopipe glass profile CAE design method of claim 1, wherein, In step S2, the runabaqus function parameters path is the program running path, InpFile is the iteration inp file to be submitted, and cpus is the number of computer cores occupied by Abaqus (2~8 cores).

4. The isopipe glass profile CAE design method of claim 1, wherein, In step S4, the center point difference is the Euclidean distance between the average coordinates of all nodes of the formed glass and the coordinates of the center point of the standard glass surface; if the center point difference is greater than the threshold and iter<Max_iter, continue iteration; otherwise, terminate the loop.

5. The isopipe glass profile CAE design method of claim 1, wherein, The method is suitable for the design of the profile of the forming die for the automobile sunroof or back glass, and is compatible with the existing design method based on the iteration compensation of the flat glass profile (the initial die profile of step S1 is replaced by the flat glass profile).

Images