Calibration process

Abstract

The present invention relates to a method for calibrating the shape of a plastically deformable component, wherein the plastically deformable component is preferably composed of a metal raw material, the method comprising: • establishing a three-dimensional theoretical model of the component; • designing a finite element model (FEM model) of the theoretical model of the component; • detecting, preferably by scanning, the three-dimensional geometric structure of the deformed actual component; • determining the deviation of the actual component relative to the theoretical model; • deforming the component by applying a force at a calculated position of the component; • checking and comparing the actual component after the deformation process with the theoretical model, wherein the force applied for the deformation and the calculated position for introducing the force at the component are determined based on the FEM model.

Description

Technical Field

[0001] The invention relates to a method for calibrating the shape of a plastically deformable component, preferably consisting of a metal raw material, comprising:

[0002] •Establish a three-dimensional theoretical model of the component;

[0003] • Finite element model (FEM model) of the theoretical model of the designed component;

[0004] • Detecting, preferably by scanning, the three-dimensional geometry of the actual deformed component;

[0005] •Determine the deviation of the actual component from the theoretical model;

[0006] • Deformation by applying forces at calculated locations on the component;

[0007] •Inspect and compare the actual component after the deformation process with the theoretical model. Background Art

[0008] A method for calibrating components is known from the prior art, in which the actual component is measured with maximum accuracy and compared to a three-dimensional theoretical model. Dimensional deviations are indicated to the worker using a computer program, who then manually corrects these areas by applying tension or pressure. To obtain an actual component that meets the dimensional requirements, these deformation and inspection processes or cycles are typically performed eight to fifteen times, which is very time-consuming.

[0009] DE 196 11 897 discloses a method for bending or calibrating metallic workpieces, in which the actual shape of the workpiece and deviations from the target shape are determined using force and distance sensors that detect jitter. During the further bending process, the determined values ​​for calculating springback and, therefore, the remaining deformation after unloading, are evaluated in a computer-assisted manner, and the bending process is terminated as soon as the calculated remaining deformation corresponds to the target shape. However, such computer-assisted methods, based on the determination of jitter using force and distance sensors, require considerable time and equipment expenditure. Summary of the Invention

[0010] The object of the present invention is to provide a method in which the calibration process of a component up to the fully calibrated component is carried out quickly and independently in such a way that as few checking and comparison processes as possible have to be carried out.

[0011] This object is achieved according to the invention in that the forces applied for the deformation and the positions calculated for introducing the forces on the component are determined on the basis of FEM.

[0012] The method according to the invention for calibrating the shape of a plastically deformable component, which preferably consists of a metal raw material, comprises the creation of a three-dimensional theoretical model of the component, preferably as a CAD model.

[0013] As a further step, the method includes designing a finite element model (FEM model) of the theoretical model of the component. This FEM model can simulate the behavior of the component under the influence of forces. Typically, such FEM models are used for dimensioning the component and designing its shape, thereby enabling the optimal shape to be determined given the prevailing forces and other influences.

[0014] The deformed actual component or its three-dimensional geometry is preferably detected by scanning, with the detected data preferably being at least temporarily stored on a computer or in the cloud. Typically, the geometry of the actual component is detected immediately following the manufacturing process. This optimizes the entire component manufacturing process, right down to component alignment. The detected data from the method or actual component, theoretical and FEM models, as established, can also preferably be collected in the cloud and incorporated as empirical values ​​for other calibration methods. Data from the manufacturing process, such as material data, casting data, heat treatment data, or other data that can be used in the method, is also incorporated into the calibration process. This data is preferably collected in the cloud and can be accessed and intercalated by different systems.

[0015] The detected data allows the determination of deviations of the actual component from the theoretical model. Based on the detected three-dimensional geometry of the actual component and the created theoretical model, the shape deviations can be determined. Preferably, a computer program is used for this purpose that allows for rapid implementation and adjustment of the detected, as-created, data.

[0016] Based on the determined deviations, a defined deformation can be implemented by applying forces at the calculated positions of the actual component. The shape deviations, based on which the deformation can be implemented, can be displayed optically by the system.

[0017] After the deformation is performed, the actual component is checked and compared with the theoretical model. If the actual component does not reach the shape of the theoretical model, the deformation process is repeated based on the newly determined deviations between the actual component and the theoretical model. Subsequently, the actual component is also checked and compared with the theoretical model again, i.e., the cycle is repeated. These processes or cycles are repeated as follows until the actual component corresponds to the theoretical model.

[0018] According to the present invention, the method is optimized by determining the forces to be applied for deformation and the calculated positions for introducing the forces on the actual component based on the FEM model, or by forming the basis for the first deformation. Preferably, the system or computer predefines defined forces based on the FEM model, which are introduced at the corresponding calculated positions to achieve the calculated deformation of the actual component. By predetermining the defined forces acting on the actual component and also the precise positions at which the forces are to be introduced, the deformation occurring in the actual component can be predefined using the FEM model. This reduces the number of cycles because, by determining the expected deformation based on the predefined forces and positions for introducing the forces using the FEM model, a real component that approximates the theoretical model is already obtained after the first cycle.

[0019] Advantageously, the FEM model is used as the basis for a first calibration process, which is carried out based on the FEM model and which allows relatively accurate determination of the deformations to be expected in the actual component using the stored FEM model.

[0020] To optimize the results or calibration process, after each deformation cycle and subsequent checking and comparison, the data collected, such as the applied force, position, and achieved deformation, are recorded by the system or computer. The forces and positions are then optimized, applied in the next cycle, and the optimized forces and positions are applied from the outset for series components. In the best case, a single deformation cycle is sufficient to calibrate the actual component so that it corresponds to the theoretical model. This data can be stored in the cloud and exchanged between systems.

[0021] For this optimization, the forces applied to deform the component and the calculated positions are preferably determined using an algorithm. That is, after each deformation (where the forces and positions are predefined and then checked and compared with the actual component), the data is recorded and optimized in a system or computer using an algorithm, so that the more cycles are performed or the more actual components are calibrated, the fewer cycles are required, as the method is optimized using the algorithm or artificial intelligence. While ideally, one cycle would still be required, empirical data have shown that a reduction of three to six cycles is significantly better than current state-of-the-art technology, which can only perform cycles less than eight.

[0022] The algorithm for determining the forces and the positions for applying them is preferably optimized using newly acquired data after each inspection and comparison of the actual model with the three-dimensional theoretical model. The newly acquired data indicates the effective deformation of the actual component with the forces applied at defined locations. Based on this newly acquired data, the basic data for the FEM model used in the first cycle can be optimized and refined. This refines and optimizes the data required for the ideal deformation each time, thereby reducing cycles and shortening process times. The method is thus self-learning and self-optimizing, or the algorithm itself is optimized.

[0023] Advantageously, the algorithm for determining the forces and the positions for introducing them is optimized using the collected data after each inspection and comparison of the actual model with the three-dimensional theoretical model. The collected data preferably consists of newly acquired data from the respective cycle. Advantageously, additional data about the actual component is also incorporated, originating from previous processes, such as the manufacturing process or any heat treatment. All of this data can be collected in the cloud and made available for optimization of the method or algorithm according to the present invention.

[0024] The present invention is also distinguished by the application of force to the actual component using a robot. By applying force to the actual component using a robot, the force or force magnitude can be accurately applied and implemented. Furthermore, the robot can be monitored during application to determine whether accurate application is being achieved. Preferably, data can also be acquired by the system, which can then be processed within the system using algorithms to optimize the achieved deformation.

[0025] Advantageously, the forces are applied to the actual component as tension and compression. This enables a specific deformation of the actual component in order to obtain the desired shape of the theoretical component.

[0026] The method according to the present invention is preferably applied to structural components from automotive engineering, as these are generally thin-walled and therefore prone to distortion during production and must subsequently be calibrated. Die-cast components are particularly preferred for this method, as these thin-walled components, typically made of light metal, distort during casting or forming and must subsequently be calibrated. Furthermore, the high production volumes of such components make such process optimization economically viable.

[0027] All design possibilities can be freely combined with one another. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] An exemplary embodiment of the present invention is described with reference to the accompanying drawings, without the invention being restricted to these exemplary embodiments.

[0029] Figure 1 A flow chart showing the sequence of the method according to the present invention. DETAILED DESCRIPTION

[0030] The actual component or its three-dimensional geometry is preferably digitally detected using a scanned geometry. This data is stored and compared with a theoretical model that was created digitally, preferably as a CAD model. This comparison reveals any deviations in the geometries. Based on the FEM model stored in the system (which was created based on the three-dimensional theoretical model) and the deviations between the actual component and the theoretical model, it is possible to determine which forces must be applied at which locations to achieve the theoretical model geometry. The system then deforms the actual component based on the calculated forces and positions for applying these forces. In the first cycle, these forces and positions are preferably based solely on the FEM model. In each subsequent cycle (if the method is performed on the same component or components of the same shape), these data, and thus the applied forces and positions, are optimized using empirical values ​​from previously calibrated components. That is, after the first deformation process, the actual component is compared with the theoretical model and the data is detected. This data is stored in the system or computer, or also in the cloud, and is preferably supplemented with other data, such as material data. If the actual component's target shape is not reached, it undergoes further cycles. Deviations from the target model are then determined, and deformation is repeated using newly acquired data and data collected from earlier cycles and present in the system. This deformation is determined using an algorithm that is optimized after each cycle, thus forming a self-learning method. The repetition of the cycle ends when the actual component corresponds to the target model.

Claims

1. A method for calibrating the shape of a plastically deformable component, the plastically deformable component being composed of a metal raw material, the method comprising: Establish a three-dimensional theoretical model of the component; Design a finite element model (FEM model) of the theoretical model of the component; Detect the 3D geometry of the actual deformed component by scanning; Determining the deviation of the actual component from the theoretical model; Deforming by applying a force at the calculated position of the component; Checking and comparing the actual component with the theoretical model after the deformation process, It is characterized in that the force applied for the deformation and the position calculated for introducing the force on the component are determined based on the FEM model.

2. The method according to claim 1, characterized in that The FEM model serves as the basis for a first calibration process.

3. The method according to claim 1, characterized in that The force applied to deform the component and the calculated position are determined with the aid of an algorithm.

4. The method according to claim 3, characterized in that After each checking and comparison process of the actual model with the three-dimensional theoretical model, the algorithm is optimized using the newly acquired data.

5. The method according to claim 4, characterized in that The algorithm is optimized with the aid of the collected data after each checking and comparison process of the actual model with the three-dimensional theoretical model.

6. The method according to any one of claims 1 to 5, characterized in that The force is applied to the actual component by means of a robot.

7. The method according to any one of claims 1 to 5, characterized in that The forces are applied to the actual component as tension and compression.

8. The method according to any one of claims 1 to 5, characterized in that The component is thin-walled.

9. The method according to any one of claims 1 to 5, characterized in that The components are structural parts from vehicle technology.

10. The method according to any one of claims 1 to 5, characterized in that The component is a die-cast part.

Images