Adaptive generation method of finite element model for FFF thermal stress analysis

Abstract

The present invention discloses a method for adaptively generating a finite element model for FFF thermal stress analysis. The method comprises the following steps: Step 1: generating a multi-resolution voxelized model of a geometric model to be simulated; Step 2: configuring the attributes of the voxelized model; Step 3: adaptively generating a multi-resolution finite element model based on the attribute configuration and the multi-resolution voxel model; Step 4: solving the generated finite element model. The present invention adopts a multi-resolution voxelization method to realize the adaptive generation of an integrated finite element model, thereby realizing the simulation reproduction of the FFF process. First, the boundary representation model is preprocessed to obtain its physical properties. By generating a multi-resolution voxel model, a hierarchical relationship and an attribute transfer relationship between them are established, and a finite element model of thermal stress analysis of the corresponding resolution can be adaptively generated, thereby improving the efficiency of model construction and enhancing the reusability and universality of the finite element model.

Description

Technical Field

[0001] The invention belongs to the field of finite element simulation analysis, and in particular relates to an integrated finite element model adaptive generation method for FFF thermal stress analysis. Background Art

[0002] FFF (Fused Filament Fabrication) is a relatively mature and most widely used additive manufacturing technology on the market. The process principle of this technology is that the material is heated to a molten state through the nozzle, and the nozzle moves along the contour and internal structure of the model under the command of the control system. The molten material is extruded while moving, and the material quickly solidifies to complete the deposition at the current position. The time consumed in this process is usually closely related to the size of the model, the area of ​​the overhanging area, the material filling rate, etc. The larger the model, the more overhanging areas, and the longer the printing time. In order to determine whether the model to be printed can be effectively printed before the actual printing, finite element analysis technology is widely used, mainly to address the problem of reduced part molding accuracy due to uneven shrinkage or curling of the material.

[0003] Before performing finite element simulation analysis on the CAD model to be printed, the corresponding finite element (analysis) model of the model is established, which usually includes a number of preliminary preparations, such as meshing; setting parameters such as material properties, ambient temperature, number of time steps and step length; adding thermal convection and constraints; loading mobile heat sources to the corresponding mesh surface to simulate nozzles; setting "unit life and death" properties to simulate the material deposition process, etc. The above preparations are currently mainly completed by manual interaction, which is time-consuming and inefficient. At the same time, the above manual efforts will not only increase with the increase in the complexity of the model, but also increase according to the increase in the accuracy requirements of the analysis (i.e., using higher resolution finite element meshes). This seriously affects the widespread use of finite element simulation analysis in FFF.

[0004] In addition, existing CAD models for FFF finite element analysis usually cannot include hole features. This is mainly because such structures often correspond to the printer nozzle running idle, and its simulation time is difficult to obtain accurately, which will cause the heat source movement and the activation of the life and death unit during the simulation process to be unable to be carried out effectively, resulting in poor simulation results.

[0005] In view of the above problems, the present invention proposes an integrated finite element model adaptive generation method for FFF thermal stress analysis. By establishing a multi-resolution voxel model and establishing transfer rules for its physical properties, efficient and convenient multi-resolution finite element model adaptive generation and reconstruction are achieved. The bounding box of the model is divided into voxels of different resolutions, and the "idle driving" and "deposition" states of the nozzle are simulated by distinguishing the voxel types, so that it can simulate the printing process of parts with arbitrary geometric shapes.

[0006] Jaiswal P, Rai RA Geometric reasoning approach for additive manufacturing print quality assessment and automated model correction[J]. Computer-Aided Design, 2018, 109. Summary of the invention

[0007] The main purpose of the present invention is to provide an integrated finite element model adaptive generation method for FFF thermal stress analysis, so as to solve the problem that it is time-consuming and laborious to manually establish a simulation model, and the establishment of a geometric model is independent of the establishment of a corresponding FFF thermal stress analysis finite element model, thereby improving the reusability and universality of the simulation model. By generating a multi-resolution voxel model, the geometric content of the model is obtained, and then the physical properties of each unit are obtained, and the finite element model is automatically constructed based on this. And by transferring physical properties between voxel models of different resolutions, a multi-resolution finite element model is adaptively generated.

[0008] The finite element model adaptive generation method for FFF thermal stress analysis includes the following steps:

[0009] Step 1: Generate a multi-resolution voxelized model of the geometric model to be simulated;

[0010] Considering that in the process of filament deposition of the printer nozzle based on FFF, the shape of the filament per unit time can be approximated as a cube, and it is convenient to unify the shapes of all finite element grid units of the model, the present invention adopts a voxelized model to further construct a finite element model. In order to realize the generation of an integrated multi-resolution finite element model, multiple resolutions are set to voxelize the geometric model to be simulated. In addition, a voxel is established according to the AABB bounding box of the model as a voxel model. After obtaining the multi-resolution voxel model, a hierarchical relationship is established for it: the first layer is the root node, which is used to represent the model bounding box voxel model, which belongs to the bottom voxel model. The second layer represents voxel models of other resolutions. The third layer is the leaf node, which represents the highest resolution voxel model and is at the highest level voxel model.

[0011] Step 2: Configure the attributes of the voxelized model;

[0012] According to the order of the printing path, the voxel units in each multi-resolution voxel model are numbered according to the order in which they are generated. In order to set the parameters of the finite element model, a resolution voxel model is selected to obtain the semantics of each voxel, that is, the physical properties. The physical properties of the finite element model for FFF include: voxel type, time step properties, "unit life and death" properties, material properties, constraints, mobile heat sources and their loading surfaces. The generated physical properties can be transferred between multi-level voxel models.

[0013] Step 3: Adaptively generate a multi-resolution finite element model based on the attribute configuration and the multi-resolution voxel model;

[0014] Establish the transfer relationship and numbering relationship of the physical properties between voxel models of different resolutions. If the physical properties of the voxel model of the required resolution have been obtained in step 2, directly import them into the finite element model. If the physical properties of the voxel model of the required resolution are unknown, obtain the physical properties corresponding to the voxel model of the required resolution based on the known physical properties and transfer relationship.

[0015] Using simulation software, an empty temperature-structure coupling model is established, and the geometric model and its physical properties are imported using a scripting language to establish a finite element model of the current resolution.

[0016] Step 4: Solve the generated finite element model.

[0017] According to the requirements, add the required result objects to the solver of the finite element model and solve to obtain the results.

[0018] The beneficial effects of the present invention are as follows:

[0019] The present invention aims at the problem that the current manual establishment of finite element models is time-consuming and laborious, and the establishment of boundary representation models is independent of the establishment of corresponding FFF thermal stress analysis finite element models. The present invention adopts a multi-resolution voxelization method to realize the adaptive generation of an integrated finite element model. First, the boundary representation model is preprocessed to obtain its physical properties. By generating a multi-resolution voxel model, the hierarchical relationship and attribute transfer relationship between them are established, and the finite element model of thermal stress analysis of the corresponding resolution can be adaptively generated, thereby improving the model construction efficiency and enhancing the reusability and universality of the finite element model. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of the present invention;

[0021] Figure 2 A boundary representation model for an embodiment;

[0022] Figure 3 is the highest resolution voxel model of the embodiment;

[0023] Figure 4 A schematic diagram of the hierarchical relationship of a multi-resolution voxel model;

[0024] Figure 5 is a schematic diagram of voxel unit types;

[0025] Figure 6 Schematic diagram of the voxel correspondence between different resolutions;

[0026] Figure 7 The simulation results of the finite element model with a voxel size of 0.4 mm;

[0027] Figure 8 The simulation results of the finite element model with a voxel size of 0.2 mm. DETAILED DESCRIPTION

[0028] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0029] like Figure 1 As shown in the figure, the automatic and rapid generation method of finite element model for FFF thermal stress analysis mainly includes the following three steps:

[0030] Step 1: Generate a multi-resolution voxelized model that is adaptive to the geometric model to be simulated.

[0031] 1-1. Determine the highest resolution.

[0032] The highest resolution means the smallest voxel size. The printable diameter of the model's failure-prone area is used as the voxel size at the highest resolution to voxelize the model. If the model does not have a failure-prone area, 0.2 mm is used as the voxel size at the highest resolution.

[0033] According to the identification method of failure-prone areas proposed by Jaiswal, the failure-prone areas are first defined as thin-walled structures, sharp-angle structures, small holes and narrow gaps.

[0034] Thin-walled structures are those where the wall thickness is less than the nozzle diameter (0.4mm). Thin-walled structures require too high a printing resolution, which can cause errors in the printing process. By obtaining the wall thickness of each part of the model and judging whether the wall thickness of each part is less than the nozzle diameter, thin-walled structures can be determined.

[0035] Sharp corners are sharp corners in the model. Due to the limited resolution of the printer, it is difficult to print sharp corners. The angle between adjacent edges in each slice layer is obtained to determine whether the area is a sharp corner structure.

[0036] The reason why small holes and narrow gaps are difficult to print is also the limited resolution of the printer. When too much or too little material is deposited, the outline of the small holes and narrow gaps will be distorted or even disappear. Small holes and narrow gaps are identified through morphological corrosion and expansion methods.

[0037] After determining the failure-prone area, the size that can successfully print the area (mainly thin-walled structures, small holes and narrow gaps) is used as the voxel size of the highest resolution voxel model. Import the geometry model to be simulated into SpaceClaim (boundary representation model, Figure 2 The boundary representation model of the embodiment) is voxelized according to the set size of the AABB bounding box of the model, such as Figure 3 The finite element model established based on the voxel model of this resolution can analyze the effect of successful printing of the model. If the finite element model is established based on a voxel model with a lower resolution, the validity of the simulation results cannot be guaranteed.

[0038] 1-2. Determine other resolutions and generate corresponding voxel models.

[0039] In order to achieve FFF simulation under different printing precisions, the common nozzle diameters of 0.4mm, 0.6mm, and 0.8mm are used as the voxel sizes of other resolution voxel models, and the models are voxelized to generate voxel models corresponding to the printing precision. It should be noted that the voxel size of other resolution voxel models is an integer multiple of the voxel size of the highest resolution voxel model. In addition, a voxel is created according to the AABB bounding box of the model as the model bounding box voxel model.

[0040] 1-3. Construct the hierarchical relationship of voxel model.

[0041] The multi-resolution voxel model corresponds to a hierarchical relationship, represented by a tree structure, such as Figure 4 As shown in the figure (the voxel size of the second-layer voxel model is 0.4mm as an example), each node in the tree represents a voxel of its corresponding voxel model. The first layer is the root node, which is used to represent the model bounding box voxel model and belongs to the bottom voxel model. The second layer represents other resolution voxel models. The third layer is the leaf node, which represents the highest resolution voxel model and is at the highest level voxel model. According to this hierarchical relationship, the transfer rules of physical properties between layers can be established, thereby establishing a finite element model. Since the first layer is the model bounding box voxel model, which is composed of only one voxel, its physical properties and the propagation relationship with other layers are not considered.

[0042] Step 2: Configure the attributes of the voxelized model.

[0043] 2-1. Number the voxels according to the printing path.

[0044] According to the order of the printing path, the voxels in each multi-resolution voxel model are numbered according to the order in which they are generated. This operation facilitates the later definition of the geometric structure and state of the "unit life and death" attribute. In the example, the unidirectional printing direction of the positive direction of the X axis is taken as an example.

[0045] 2-2. Obtain the physical properties of each voxel in a high-resolution voxel model.

[0046] A voxel in the multi-resolution voxel model corresponds to a solid unit in the finite element model, and a solid unit is activated in one unit time. In order to set the parameters of the finite element model, it is necessary to obtain the semantics of each voxel in a voxel model in advance, that is, the physical properties. The voxel model is the highest resolution voxel model, or other resolution voxel models determined according to the actual printer nozzle diameter.

[0047] The physical properties of voxels include: voxel type, time step properties, "unit life and death" properties, material properties, constraints, moving heat sources and their loading surfaces.

[0048] (1) Voxel type:

[0049] There are three types of voxel: Figure 5 As shown:

[0050] Model internal voxel: A voxel whose interior or surface points are all located inside the model (excluding the model surface) is called a model internal voxel.

[0051] Model boundary voxel: A voxel that intersects with the outer surface of the model (i.e., the interior or surface of the voxel contains at least one point located on the outer surface of the model) is called a model boundary voxel.

[0052] Empty voxel: The eight vertices of a voxel are all outside the model and do not intersect with the model. This type of voxel is called an empty voxel.

[0053] (2) Time step attributes:

[0054] The time step attributes of a voxel include the total time step, step length, and activation time step, which depend on the last number of the voxel model (i.e., the total number of voxels), the nozzle movement speed, the voxel size, and the current voxel number. The total time step is the total number of the voxel model + 1, where 1 is the time step of the cooling process. Each voxel in the same resolution voxel model has the same total time step. The step length is the generation time of a voxel, which is determined by dividing the side length of a voxel by the nozzle movement speed. The activation time step is the current voxel number, which means that all voxels before this number have been activated. Among them, the activation of voxels inside the model and voxels at the boundary of the model represents the generation process of the voxel, and the activation of empty voxels represents the idle driving process of the nozzle.

[0055] (3) “Unit life and death” attribute:

[0056] The "unit life and death" attribute of the voxel corresponds to the time step attribute and is used to simulate the deposition process of FFF. The "unit life and death" attribute includes two states: "inhibited" state and "activated" state, which corresponds to a certain state at each time step. The "inhibited" state means that the unit has not been deposited, and the "activated" state means that the unit has been generated. Because in the FFF process, once the unit is deposited, it means that the printing of the position is completed. Therefore, the initial state of the "unit life and death" attribute of each voxel unit is "inhibited". According to the activation time step, the states corresponding to the activation time step and all subsequent time steps are converted to "activated" to indicate the generation of the unit.

[0057] (4) Material properties:

[0058] The material properties of the voxel unit are the physical properties of the raw materials used for printing, which are customized by the user according to their own needs. They usually include specific heat capacity, density, elastic modulus, Poisson's ratio, thermal expansion coefficient, etc.

[0059] (5) Constraints:

[0060] Constraints refer to the support provided by the printing plate and support structure to the model during the printing process. Therefore, for the voxels that are attached to the printing plate and support structure, it is necessary to impose "fixed support" constraints on their lower surfaces. Generally, only the boundary voxels of the model have "fixed support" constraints.

[0061] (6) Mobile heat source and its loading surface:

[0062] The activation process of the voxel unit is triggered by the action of the moving heat source on its upper surface, and the loading surface is the upper surface of each layer of voxels. For the simulation of the FFF process, the embodiment of the present invention adopts (but is not limited to) a Gaussian heat source to establish a nozzle heat source model. Since the printing direction of the embodiment is the positive printing of the X-axis, the heat source moves parallel to the XY plane along the positive direction of the X-axis, and the formula is:

[0063]

[0064] Where q is the calorific value of the coordinate point, Q m is the maximum heat flux density at the center of the heat source, v is the moving speed of the heat source, t is the time, and R is the radius of the heat source.

[0065] Step 3: Based on the attribute configuration and the multi-resolution voxel model, adaptively generate a multi-resolution finite element model.

[0066] 3-1. Determine the relationship between the numbers of multi-resolution voxel models:

[0067] The relationship between the numbers of two different resolution voxel models with voxel sizes of P (other resolution) and T (highest resolution) is as follows, where the voxel sizes between different resolutions are N (P / T, which is an integer by default). The bounding box size of the part is X×Y×Z, so the specifications of the P resolution voxel model are The specifications of the T-resolution voxel model are

[0068] First, for any resolution voxel model (S is the voxel size), the relationship between its voxel number R and the corresponding position in the model (rth layer, t row, v column) is as follows:

[0069]

[0070]

[0071] It is known that a voxel number in the P voxel model is I. According to formula (1), its corresponding position in the voxel model is the i-th layer, j-th row, and k-th column. This voxel corresponds to N in the T voxel model. 3 Voxels (these voxels are a cube). Through voxel I, the position of its corresponding voxel in the T voxel model can be determined. Figure 6 As shown, the voxel in the 1st layer, 1st row, and 1st column of the cube corresponds to the i1th layer, j1th row, and k1th column in the T voxel model, and the voxel in the Nth layer, Nth row, and Nth column of the cube corresponds to the i2th layer, j2th row, and k2th column in the T voxel model, where:

[0072]

[0073] i2=i1+N-1,j2=j1+N-1,k2=k1+N-1, Formula (4)

[0074] Then, according to its position, its number is obtained by formula (2):

[0075] On the contrary, if it is known that there is a voxel in the T voxel model, whose number is M, and its position is calculated according to formula (1) to be in the qth layer, mth row, and nth column, then its corresponding voxel position in the P voxel model is q'th layer, m'th row, and n'th column:

[0076]

[0077] The number can be obtained by formula (2).

[0078] 3-2. Establishing the rules for transferring physical properties between multi-level voxel models:

[0079] By establishing a corresponding relationship between voxel numbers between multi-resolution voxel models, the physical properties between related voxels can be transferred. The physical properties that can be transferred are as follows:

[0080] (1) Voxel type:

[0081] For different levels of voxel type transfer, if the high-level voxel model contains model boundary voxels, the corresponding low-level voxel type is model boundary voxel; if the high-level voxel model only contains model internal voxels, the corresponding low-level voxel model type is model internal voxel; if the high-level voxel model only contains empty voxels, then the corresponding low-level voxel model is empty voxel.

[0082] (2) Time step attributes:

[0083] For voxel models at different levels, the relationship between their step sizes can be expressed as L = N 3 l, where L is the step size of the low-level voxel model, l is the step size of the high-level voxel model, and N is the ratio of the voxel size of a low-level voxel model to that of the high-level voxel model.

[0084] (3) Material properties:

[0085] For each level of voxel model, its material properties can be transferred to each other, that is, the material properties of the voxels in the low-level voxel model are the material properties of the corresponding voxels in the high-level voxel model, and vice versa.

[0086] (4) Constraints:

[0087] For the transmission of constraint conditions, if there is a "fixed support" constraint on the lower surface in the boundary voxels of the high-level model, the lower surface of the voxels in the corresponding low-level voxel model also has this constraint. Conversely, if there is a lower surface constraint in the boundary voxels of the low-level model, the corresponding voxels in the high-level voxel model only have the same "fixed support" constraint on the lower surface of the bottom layer of voxels.

[0088] (5) Mobile heat source and its loading surface:

[0089] The moving heat sources all obey the same heat source formula. The loading surface of the moving heat source is similar to the constraint conditions. When the upper surface of the voxel in the high-level model is the loading surface, the upper surface of the corresponding low-level voxel is also the loading surface; conversely, when the upper surface of the low-level voxel is the loading surface, the voxel corresponding to the high-level voxel model only has the upper surface of the topmost layer of voxels as the loading surface.

[0090] 3-3. Generate a finite element model of the required resolution.

[0091] If the required resolution is the resolution of the voxel model of the physical properties obtained in step 2, the physical properties are known; if not, the physical properties of the voxel model of the required resolution are obtained based on the obtained physical properties and their transfer relationship.

[0092] Use simulation software, in this case Ansys Workbench, to build a finite element model. Use the AdditiveWizard plug-in to build an empty temperature-structure coupling model, and use a script file to automatically import the geometric model and its physical properties into the empty model. The finite element model of the required resolution is then built.

[0093] Step 4: Solve the generated finite element model.

[0094] Add the required result objects to the solver of the generated finite element model and solve it.

[0095] like Figure 7 The figure shows the temperature field and total deformation results of the finite element model (at a certain moment) with a voxel size of 0.4 mm. Figure 8 Shown are the temperature field and total deformation results of the finite element model (at a certain moment) with a voxel size of 0.2 mm.

Claims

1. A finite element model adaptive generation method for FFF thermal stress analysis, characterized in that: The following steps are involved: Step 1: Generate a multi-resolution voxelized model of the geometric model to be simulated; The voxelized model is used to further construct the finite element model, and multiple resolutions are set to voxelize the geometric model to be simulated; in addition, a voxel is established according to the AABB bounding box of the model as a voxel model; after obtaining the multi-resolution voxel model, a hierarchical relationship is established for it: the first layer is the root node, which is used to represent the model bounding box voxel model and belongs to the bottom voxel model; the second layer represents other resolution voxel models; the third layer is the leaf node, which represents the highest resolution voxel model and is at the highest level voxel model; Step 2: Configure the attributes of the voxelized model; According to the order of the printing path, the voxel units in each multi-resolution voxel model are numbered according to the order in which they are generated; in order to set the parameters of the finite element model, a resolution voxel model is selected to obtain the semantics of each voxel, that is, the physical properties; the physical properties of the finite element model for FFF include: voxel type, time step properties, "unit life and death" properties, material properties, constraints, mobile heat sources and their loading surfaces; the generated physical properties can be transferred between multi-level voxel models; Step 3: Adaptively generate a multi-resolution finite element model based on the attribute configuration and the multi-resolution voxel model; Establish the transfer relationship and numbering relationship of physical properties between voxel models with different resolutions; if the physical properties of the voxel model of the required resolution have been obtained in step 2, directly import them into the finite element model; if the physical properties of the voxel model of the required resolution are still unknown, obtain the physical properties corresponding to the voxel model of the required resolution based on the known physical properties and transfer relationship; Using simulation software, establish an empty temperature-structure coupling model, import the geometric model and its physical properties to establish a finite element model of the current resolution; Step 4: Solve the generated finite element model; According to the requirements, add the required result objects to the solver of the finite element model and solve to obtain the results.

2. The method for adaptively generating a finite element model for FFF thermal stress analysis according to claim 1, characterized in that: Step 1: Generate a multi-resolution voxelized model that is adaptive to the geometric model to be simulated; 1-1. Determine the highest resolution; The highest resolution means the smallest voxel size. The printable diameter of the failure-prone area of ​​the model is used as the voxel size at the highest resolution to voxelize the model. If the model does not have a failure-prone area, 0.2 mm is used as the voxel size of the highest resolution; According to the identification method of failure-prone areas proposed by Jaiswal, the failure-prone areas are first defined as thin-walled structures, sharp-angle structures, small holes, and narrow gaps; Thin-walled structures are those where the wall thickness is smaller than the nozzle diameter. Thin-walled structures require too high a printing resolution, which can cause errors in the printing process. By obtaining the wall thickness of each part of the model, it is determined whether the wall thickness of each part is less than the nozzle diameter to determine the thin-wall structure; Sharp corners are sharp corners in the model. Due to the limited resolution of the printer, it is difficult to print sharp corners. By obtaining the angle between adjacent edges in each slice, it is determined whether the area is a sharp-angle structure; The reason why small holes and narrow gaps are difficult to print is also due to the limited resolution of the printer; when too much or too little material is deposited, the outlines of small holes and narrow gaps will be distorted or even disappear; small holes and narrow gaps are identified through morphological corrosion and expansion methods; After determining the area prone to failure, the size that enables successful printing of the area is used as the voxel size of the highest resolution voxel model; import the geometric model to be simulated into SpaceClaim, and voxelize the AABB bounding box of the model according to the set size; 1-2. Determine other resolutions and generate corresponding voxel models; In order to realize FFF simulation under different printing precisions, the nozzle diameters of 0.4mm, 0.6mm, and 0.8mm are used as the voxel sizes of other resolution voxel models, and the models are voxelized respectively to generate voxel models corresponding to the printing precisions. It should be noted that the voxel sizes of other resolution voxel models are integer multiples of the voxel sizes of the highest resolution voxel models. In addition, a voxel is created according to the AABB bounding box of the model as the model bounding box voxel model. 1-3. Construct the hierarchical relationship of the voxel model; The multi-resolution voxel model corresponds to a hierarchical relationship, which is represented by a tree structure. Each node in the tree represents a voxel of the corresponding voxel model. The first layer is the root node, which is used to represent the model bounding box voxel model and belongs to the bottom voxel model. The second layer represents other resolution voxel models. The third layer is the leaf node, which represents the highest resolution voxel model and is at the highest level voxel model. According to this hierarchical relationship, the transmission rules of physical properties between layers can be established, thereby establishing a finite element model. Since the first layer is the model bounding box voxel model, which is composed of only one voxel, its physical properties and the transmission relationship with other layers are not considered.

3. The method for adaptively generating a finite element model for FFF thermal stress analysis according to claim 2, characterized in that: Step 2: 2-1. Number the voxels according to the printing path; According to the order of the printing path, the voxels in each multi-resolution voxel model are numbered according to the order in which they are generated; 2-2. Obtain the physical properties of each voxel in a resolution voxel model; A voxel in the multi-resolution voxel model corresponds to a solid unit in the finite element model, and a solid unit is activated in one unit time. In order to set the parameters of the finite element model, it is necessary to obtain the semantics of each voxel in a voxel model in advance, that is, the physical properties. The voxel model is the highest resolution voxel model, or other resolution voxel models determined according to the actual printer nozzle diameter. The physical properties of voxels include: voxel type, time step properties, "unit life and death" properties, material properties, constraints, moving heat sources and their loading surfaces; (1) Voxel type: There are three types of voxels: Model internal voxels: Voxels whose interior or surface points are all located inside the model are called model internal voxels; Model boundary voxels: voxels intersect with the outer surface of the model, that is, the voxel interior or surface contains at least one point located on the outer surface of the model, which is called a model boundary voxel; Empty voxel: The eight vertices of the voxel are all outside the model and do not intersect with the model. This type of voxel is called an empty voxel; (2) Time step attributes: The time step attributes of a voxel include the total time step, step length and activation time step, which depend on the last number of the voxel model, i.e. the total number of voxels, the nozzle movement speed, the voxel size and the current voxel number; the total time step is the total number of the voxel model + 1, where 1 is the time step of the cooling process; each voxel in the same resolution voxel model has the same total time step; the step length is the generation time of a voxel, which is determined by dividing the side length of a voxel by the nozzle movement speed; the activation time step is the current voxel number, which means that all voxels before this number have been activated; among them, the activation of the voxels inside the model and the voxels at the boundary of the model represents the generation process of the voxel, and the activation of the empty voxels represents the empty driving process of the nozzle; (3) "Unit life and death" attribute: The "unit life and death" attribute of the voxel corresponds to the time step attribute and is used to simulate the deposition process of FFF; the "unit life and death" attribute includes two states: "inhibited" state and "activated" state, which corresponds to a certain state at each time step; the "inhibited" state means that the unit has not been deposited, and the "activated" state means that the unit has been generated; in the FFF process, once the unit is deposited, it means that the voxel position is printed; therefore, the initial state of the "unit life and death" attribute of each voxel unit is "inhibited", and according to the activation time step, the states corresponding to the activation time step and all subsequent time steps are converted to "activated" to indicate the generation of the unit; (4) Material properties: The material properties of the voxel unit are the physical properties of the raw materials used for printing, which can be customized by the user according to their own needs; (5) Constraints: Constraints refer to the support provided by the printing plate and support structure to the model during the printing process. Therefore, for the voxels that are attached to the printing plate and support structure, it is necessary to impose "fixed support" constraints on their lower surfaces. (6) Mobile heat source and its loading surface: The activation process of the voxel unit is triggered by the action of a moving heat source on its upper surface, and the loading surface is the upper surface of each layer of voxels.

4. The method for adaptively generating a finite element model for FFF thermal stress analysis according to claim 3, characterized in that: In the simulation of the FFF process in the moving heat source and its loading surface, a Gaussian heat source is used to establish the nozzle heat source model. When the printing direction is the positive printing direction of the X-axis, the heat source moves parallel to the XY plane along the positive direction of the X-axis. The formula is: Where q is the calorific value of the coordinate point, Q m is the maximum heat flux density at the center of the heat source, v is the moving speed of the heat source, t is the time, and R is the radius of the heat source.

5. The method for adaptively generating a finite element model for FFF thermal stress analysis according to claim 3 or 4, characterized in that: Step 3: 3-1. Determine the relationship between the numbers of multi-resolution voxel models: The relationship between the numbers of two different resolution voxel models with voxel sizes of P and T is as follows, where the voxel sizes of different resolutions are N times; the bounding box size of the part is X×Y×Z, then the specifications of the P resolution voxel model are The specifications of the T-resolution voxel model are First, for any resolution voxel model, S is the voxel size, and its voxel number R is related to the rth layer, tth row, and vth column of the corresponding position in the model as follows: It is known that a voxel number in the P voxel model is I. According to formula (1), its corresponding position in the voxel model is the i-th layer, j-th row, and k-th column. This voxel corresponds to N in the T voxel model. 3 voxel; through voxel I, the position of its corresponding voxel in the T voxel model can be determined; the voxel in the 1st layer, 1st row, and 1st column of the cube corresponds to the i1th layer, j1th row, and k1th column in the T voxel model, and the voxel in the Nth layer, Nth row, and Nth column of the cube corresponds to the i2th layer, j2th row, and k2th column in the T voxel model, where: i2=i1+N-1,j2=j1+N-1,k2=k1+N-1, Formula (4) Then, according to its position, its number is obtained by formula (2): On the contrary, if it is known that there is a voxel in the T voxel model, whose number is M, and its position is calculated according to formula (1) to be in the qth layer, mth row, and nth column, then its corresponding voxel position in the P voxel model is q'th layer, m'th row, and n'th column: Then the number can be obtained by formula (2); 3-2. Establishing the rules for transferring physical properties between multi-level voxel models: By establishing a corresponding relationship between voxel numbers between multi-resolution voxel models, the physical properties between related voxels are transferred; the physical properties that can be transferred are as follows: (1) Voxel type: For the voxel type transfer at different levels, if the high-level voxel model contains model boundary voxels, the corresponding low-level voxel type is model boundary voxels; if the high-level voxel model only contains model internal voxels, the corresponding low-level voxel model type is model internal voxels; if the high-level voxel model only contains empty voxels, then the corresponding low-level voxel model is empty voxels; (2) Time step attributes: For voxel models at different levels, the relationship between their step sizes can be expressed as L = N 3 l, where L is the step size of the low-level voxel model, l is the step size of the high-level voxel model, and N is the ratio of the voxel size of a low-level voxel model to that of the high-level voxel model; (3) Material properties: For each level of voxel model, its material properties can be transferred to each other; that is, the material properties of the voxels in the low-level voxel model are the material properties of the corresponding voxels in the high-level voxel model, and vice versa; (4) Constraints: For the transmission of constraint conditions, if there is a "fixed support" constraint on the lower surface in the boundary voxels of the high-level model, the lower surface of the voxels in the corresponding low-level voxel model also has this constraint; conversely, if there is a lower surface constraint in the boundary voxels of the low-level model, the corresponding voxels in the high-level voxel model only have the same "fixed support" constraint on the lower surface of the bottom layer of voxels; (5) Mobile heat source and its loading surface: The moving heat sources all follow the same heat source formula. The loading surface of the moving heat source is similar to the constraint condition. When the upper surface of the voxel in the high-level model is the loading surface, the upper surface of the corresponding low-level voxel is also the loading surface. Conversely, when the upper surface of the low-level voxel is the loading surface, the voxel corresponding to the high-level voxel model has only the upper surface of the top layer of voxels as the loading surface. 3-3. Generate a finite element model with the required resolution; If the required resolution is the resolution of the voxel model of the physical properties obtained in step 2, then the physical properties are known; if not, the physical properties of the voxel model of the required resolution are obtained based on the obtained physical properties and their transfer relationship; Use the simulation software Ansys Workbench to build a finite element model; use the Additive Wizard plug-in to build an empty temperature-structure coupling model, and use the script file to automatically import the geometric model and its physical properties into the empty model; then the finite element model of the required resolution is built.

Images