CN-122021120-A - High-precision inherent strain coefficient calibration model and calibration method

CN122021120ACN 122021120 ACN122021120 ACN 122021120ACN-122021120-A

Abstract

The invention relates to the technical field of additive manufacturing, and discloses a calibration method of a high-precision inherent strain coefficient calibration model, which comprises the steps of: the topological hollowed-out structure is arranged inside the thermal conduction correction structure, the topological morphology change regulates and controls the local thermal conduction path, and the thermal history of the original inherent strain coefficient calibration model in the printing process is adjusted. And accumulating a database by acquiring the thermal history in the actual printing process of the original inherent strain coefficient calibration model in the model through data acquisition, performing supervision training on a deep neural network regression model by using the basic database, performing temperature field calculation on a complex model requiring stress simulation, and applying the corresponding inherent strain coefficient according to the thermal history of each point of the model to realize the calculation of a refined inherent strain method.

Inventors

ZHANG KEQING
Chai Pengtao
HE CHI
ZHANG GUOLIANG

Assignees

上海镭镆科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251226

Claims (10)

1. A high-precision inherent strain coefficient calibration model is characterized by comprising, An original inherent strain coefficient calibration model; the heat conduction correction structure changes the heat transfer rate and the heat dissipation path in the printing process; The topological hollowed-out structure is arranged inside the heat conduction correction structure, the topological morphology change regulates and controls the local heat conduction path, and the heat history in the printing process of the original inherent strain coefficient calibration model is regulated.
2. The high-precision inherent strain coefficient calibration model of claim 1, wherein the topological hollow structure is a lattice structure, and the density of the entire topological hollow structure is adjusted by changing the size of lattice struts.
3. The high-precision inherent strain coefficient calibration model of claim 2, wherein the filling rate of the lattice struts accounting for the whole topological hollow structure is 10% -80%, and the diameter of the lattice struts is 0.5-3mm.
4. The high precision intrinsic strain coefficient calibration model of claim 2, wherein the lattice types include, but are not limited to, cubes, face-centered cubes, body-centered cubes, tetrahedrons, cell structures of three-period extremely small curved surfaces, and composite structures thereof.
5. The high precision intrinsic strain coefficient calibration model of claim 1, wherein the raw intrinsic strain coefficient calibration model includes, but is not limited to, a cantilever beam or a cross-thin-walled structure.
6. The high precision intrinsic strain coefficient calibration model of claim 1, wherein the thermal history of different of the calibration models is obtainable by direct or indirect means.
7. A method for calibrating by using an intrinsic strain coefficient calibration model, comprising the high-precision intrinsic strain coefficient calibration model according to any one of claims 1 to 6, characterized by comprising the steps of, And S1, acquiring data related to the inherent strain coefficient calibration through an inherent strain coefficient calibration model. S2, acquiring a thermal history in an actual printing process of an original inherent strain coefficient calibration model in the model; S3, accumulating a database according to the inherent strain coefficient corresponding to the thermal history obtained in the step S2, and performing supervision training on a deep neural network regression model by using the basic database, wherein the mean square error is used as a loss function during training, and an Adam optimizer is selected; And S4, calculating a temperature field of the complex model of the required stress simulation, and applying a corresponding inherent strain coefficient according to the thermal history of each point of the model to realize the calculation of a refined inherent strain method.
8. The method of calibrating a device using an intrinsic strain coefficient calibration model according to claim 7, wherein the information obtained in step S1 includes, but is not limited to, deformation size.
9. The method of calibrating a thermal transfer process according to claim 7, wherein the obtaining in step S2 includes, but is not limited to, establishing a thermal transfer process numerical simulation model or a thermocouple.
10. The method according to claim 7, wherein in the step S3, the number of nodes of the input layer of the calibration model is consistent with the dimension of the thermal history feature data, the number of nodes of the output layer is consistent with the dimension of the intrinsic strain coefficient, and the calibration model includes at least two fully connected hidden layers.

Description

High-precision inherent strain coefficient calibration model and calibration method Technical Field The invention relates to the technical field of additive manufacturing, in particular to a high-precision inherent strain coefficient calibration model and a calibration method. Background At present, two methods, namely an inherent strain method and a thermal coupling method, are mainly adopted for finite element simulation of residual stress in the additive manufacturing process. The inherent strain method avoids the point-by-point calculation of the thermo-mechanical coupling process by equivalent complex composite strain to inherent strain, ensures certain precision and simultaneously remarkably improves calculation efficiency, and is commonly used for stress prediction of materials (such as common stainless steel, aluminum alloy and the like) subjected to heat-induced deformation. However, this approach assumes that strain is primarily due to thermal expansion and does not adequately account for the volume change caused by the phase change. The computational accuracy of conventional intrinsic strain methods is determined primarily by means of calibration experiments, which are typically performed using additive manufacturing to print only the component 3 of fig. 1, with the intrinsic strain coefficients being inversely calibrated according to the specific deformations of the component 3. Therefore, when materials with significant phase change stress (such as martensitic steel) are processed, the prediction accuracy is significantly reduced, because the dependence on the simple strain coefficient calibration model cannot reflect the stress-strain characteristics in the complex printing environment. The thermal coupling method gradually simulates thermal circulation and thermal strain, phase change strain and mechanical response caused by the thermal circulation in the printing process by directly coupling the temperature field and the structural field, and can reflect the stress evolution of the material under complex physical change more finely. However, the method has two main defects that firstly, the calculation depends on a large number of accurate boundary conditions and material parameters (especially phase change kinetic parameters, high-temperature mechanical properties and the like), the data acquisition and modeling difficulties are large, secondly, the calculation time of full coupling analysis is long, the required calculation resources are multiplied by a plurality of times compared with an inherent strain method, and the method is difficult to be suitable for a design scene of rapid iteration in engineering. The existing method is difficult to consider precision and efficiency, and particularly for phase change sensitive materials, no residual stress prediction scheme which can not only keep high-efficiency calculation but also accurately reflect the complex printing environment exists. Disclosure of Invention In order to solve the technical problems, the invention aims to provide a calibration model and an inherent strain calibration method for flexibly adjusting a thermal environment. In order to achieve the technical purpose, as a first aspect, the application is realized by a high-precision inherent strain coefficient calibration model, which comprises, An original inherent strain coefficient calibration model; the heat conduction correction structure changes the heat transfer rate and the heat dissipation path in the printing process; The topological hollowed-out structure is arranged inside the heat conduction correction structure, the topological morphology change regulates and controls the local heat conduction path, and the heat history in the printing process of the original inherent strain coefficient calibration model is regulated. According to the invention, the topological hollow structure is a lattice structure, and the density of the whole topological hollow structure is adjusted by changing the size of the lattice support. According to the invention, the filling rate of the lattice support column accounting for the whole topological hollow structure is 10% -80%, and the diameter of the lattice support column is 0.5-3mm. According to the present invention, further, the lattice types include, but are not limited to, cubes, face-centered cubes, body-centered cubes, tetrahedrons, cell structures of three-period extremely small curved surfaces, and composite structures thereof. According to the invention, further, the original intrinsic strain coefficient calibration model comprises, but is not limited to, a cantilever beam or a cross thin-wall structure. According to the invention, further, the thermal history of the different calibration models can be obtained directly or indirectly. As a second aspect, the present invention also provides a method for calibration using an intrinsic strain coefficient calibration model, comprising the high-precision intrinsic strain coefficient