A Database-Based Method for Deformation Prediction and Numerical Compensation of Additive Manufacturing Parts

Abstract

This invention discloses a database-based method for predicting and numerically compensating for deformation in additive manufacturing parts. First, the deformation amount of the printing path of the additively manufactured part is predicted using a database. Then, the original geometric model is numerically compensated based on the predicted deformation value. Finally, the compensated geometric model is additively formed along the printing path. Key technologies included in this invention are: ① the establishment and updating of the part deformation database; ② the calculation of the deformation weight coefficients of the nodes in the printing path; ③ the prediction of the deformation amount of the nodes in the printing path; and ④ the proportional reverse deformation compensation of the original geometric model of the part. This invention cleverly uses a part deformation database to predict the deformation amount of the nodes in the printing path and improves the forming accuracy of the part by changing the original geometric model through a proportional reverse deformation compensation method.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing technology, and more specifically, to a database-based method for predicting and numerically compensating for deformation of additively manufactured parts. Background Technology

[0002] Additive manufacturing integrates a range of technologies, including computer-aided design, computer-aided manufacturing, computer numerical control, and materials science, to form a target object by stacking layers of materials such as plastics, metals, and powders. The concept of additive manufacturing aims to improve traditional manufacturing processes, and its emergence has fundamentally changed the way industrial production is conducted. With the maturation of additive manufacturing technology, it has been widely applied in various fields such as aerospace, medical, construction, and aesthetic products.

[0003] The forming accuracy of additively manufactured parts directly affects their structural performance and overall quality. This accuracy is related to factors such as the printing path and forming method. Currently, there are two main methods to improve part printing accuracy: ① path planning; ② post-processing. Existing methods rely heavily on trial and error based on engineering experience, which is both inefficient and costly. Therefore, designing a standardized, efficient, and low-cost method to improve part forming accuracy is crucial. Summary of the Invention

[0004] 1. In view of the above-mentioned problems in the prior art, the present invention proposes a database-based method for predicting and numerically compensating for deformation of additive manufacturing parts.

[0005] 2. A database-based method for predicting and numerically compensating for deformation in additively manufactured parts includes the following steps:

[0006] (1) Establish the geometric model of the part and mesh the geometric model of the part. The total number of model elements is H.

[0007] (2) Obtain the node number N based on the discrete unit and the discrete precision. i ,i=1,…,n and node coordinates D={…,(x i y i , z i ),…},i=1,…,n, where (x i y i , z i ) represents the coordinates of node i.

[0008] (3) Establishment and updating of the part deformation database.

[0009] ① Randomly generate m printing paths, and initialize the k=1 (k≤m)th path.

[0010] ② Calculate the nodal strain: Based on the thermo-coupling theory simulation analysis (temperature field numerical simulation and stress field numerical simulation) of the additive manufacturing process of the k-th printing path, the strain ε={ε1,ε2,…,ε i , …, ε n}, where ε i The total dependent variable for node i.

[0011] ε i =A -1 σ i +αΔT i

[0012]

[0013] Where the total strain ε at node i i =[ε ix ε iy ε iz ε ixy ε ixy ε ixz ] T The total stress σ at node i i =[σ ix σ iy σ iz σ ixy σ ixy σ ixz ] T The coefficient of thermal expansion α = [α x α y α z 0 0 0] T E, ν, and G are Young's modulus, Poisson's ratio, and shear modulus, respectively.

[0014] ③ Calculate the nodal deformation: The nodal deformation S is derived from the nodal strain in step ②. k ={s k 1,s k 2, ..., s k i , ..., s k n}, where s k i Let s be the deformation of node i along path k. i The definition is as follows:

[0015] s k i =(s k xi s k yi s k zi)

[0016] s k xi s k yi s k zi Let be the deformation of node i along path k in the x, y, and z directions, where i = 1, ..., n.

[0017] ④ Update the database: Update the deformation of the node corresponding to path k to the database.

[0018] ⑤ Determine if k is less than or equal to m. If so, let k = k + 1 and repeat steps ②-④. If not, proceed to step (4).

[0019] (4) Calculation of deformation weight coefficients for path nodes to be printed: Initialize matching weight coefficient w k =0 (k=1,…,m; k represents the k-th printing path in the database). Check sequentially whether the T-th step (T=1,…,H-1) of the path to be printed is equal to the T-th step of each path in the database. If they are equal, let w = 0. k =w k +1.

[0020] (5) Prediction of deformation of path nodes to be printed: Predicted value s of deformation of path nodes to be printed * i Let i = 1, ..., n, which is the ratio of the sum of the products of the deformation and matching weight coefficients of the node under all paths in the database to the sum of all weight coefficients. The predicted value s of the deformation of the path node to be printed is... * i The definition is as follows:

[0021]

[0022] Where w k This represents the matching weight coefficient corresponding to the k-th printing path in the database; s k xi s k yi s k zi This represents the printing deformation of node i in the x, y, and z directions under the k-th path in the database.

[0023] (6) Compensation for the same-axis reverse deformation of the original geometric model of the part: Update the coordinates of each node, multiply the deformation of each node of the path to be printed in step (5) by (-1) in the reverse direction and add it to the node coordinates in step (2) to obtain the new coordinates of each node, i.e., the reconstructed node coordinates D'. Sweep the reconstructed node coordinates to obtain the reconstructed model. The reconstructed node coordinates D' are defined as:

[0024] D′=DS *

[0025] That is: D′={…,(x i ,yi,z i ),…}-{…,(s * xi s * yi s * zi ),...}

[0026] ={…,(x i -s * xi y i -s * yi , z i -s * zi ),...} Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of the method of the present invention;

[0029] Figure 2 Mesh model of a thin rod part;

[0030] Figure 3 Number the element and node of the thin rod part model;

[0031] Figure 4 The path to be printed;

[0032] Figure 5 This is a schematic diagram of the same-year reverse deformation compensation method;

[0033] Figure 6 To reconstruct the model using a grid;

[0034] Figure 7 Simulation results of forming a thin rod part after path deformation compensation.

[0035] Figure 8 This is the overall flowchart of the present invention; Detailed Implementation

[0036] The following are some specific implementation examples of the technical solutions adopted in this invention. It should be noted that the described embodiments are only intended to facilitate the understanding of this invention and do not limit it in any way.

[0037] Specific implementation case: Using ANSYS, we perform database-based deformation prediction and numerical compensation for slender rod parts in additive manufacturing. The deformation prediction and numerical compensation process is as follows: Figure 1 As shown.

[0038] 1. Create a geometric model of the part and mesh the geometric model of the part (e.g., Figure 2 As shown), the total number of single-layer units is 48 (as shown). Figure 3 (As shown); the node number N is obtained based on the discrete unit and the discrete precision. i ,i=1,…,1767 (such as Figure 3 (as shown) and the coordinates of each node D = {…,(x i ,yi,z i ),…},i=1,…,1767,where(x i y i , z i ) represents the coordinates of node i (as shown in Table 1).

[0039] Table 1 Node Coordinates of the Thin Rod Part Model

[0040]

[0041] 2. Establishment and updating of the part deformation database: 10,000 printing paths are randomly generated. Based on the thermo-mechanical coupling theory simulation analysis (temperature field numerical simulation and stress field numerical simulation), the additive manufacturing process of all printing paths is analyzed to obtain the strain ε = {ε1, ε2, ..., ε} of all nodes. i , …, ε 1767}, i = 1, ..., 1767, and then the nodal deformation S is calculated through the nodal strain. k ={s k 1,s k 2, ..., s k i , ..., s k n}, k = 1, ..., 10000, where s k i The deformation s of node i along path k k i =(s k xi s k yi s k zi), i = 1, ..., 1767, update the deformation of all nodes corresponding to the paths to the database (as shown in Table 2).

[0042] Table 2 Database of Nodal Deformation of Thin Rod Parts

[0043]

[0044] 3. Calculation of deformation weight coefficients for path nodes to be printed: Initialize matching weight coefficient w k =0 (k=1,…,10000; k represents the k-th print path in the database). Sequentially determine the paths to be printed (e.g., Figure 4 If the T-th step (T = 1, ..., 47) of the path shown is equal to the T-th step of each path in the database, then let w... k =w k +1.

[0045] 4. Prediction of deformation of path nodes to be printed: Calculate the predicted value of deformation of path nodes to be printed. (As shown in Table 3), where w k This represents the matching weight coefficient corresponding to the k-th printing path in the database; s k xi s k yi s k zi This represents the node deformation of node i in the x, y, and z directions under the k-th path in the database.

[0046] Table 3 Prediction of Deformation of Path Nodes to be Printed

[0047]

[0048] 5. Compensation for proportional reverse deformation of the original geometric model of the part (e.g.) Figure 5 (As shown in Table 4); Update the coordinates of each node by multiplying the deformation of each node in the path to be printed by (-1) in the reverse direction and adding it to the node coordinates to obtain the new coordinates of each node, i.e., reconstruct the node coordinates D' (as shown in Table 4). Then, obtain the reconstructed model by sweeping the reconstructed node coordinates (as shown in Table 4). Figure 6 (As shown). The reconstructed node coordinates D' are defined as: D′ = DS * ={(x i -s * xi y i -s * yi , z i -s * zi ),…}, where i=1,…,1767.

[0049] 6. Additive Manufacturing Simulation: Using the reconstructed model as the simulation object, and the path to be printed (e.g., ...) Figure 4 The image shows the laser cladding path. This is used to verify the forming accuracy of the thin rod part after deformation compensation of the printing path (e.g., ...). Figure 7 (As shown).

[0050] Table 4 Reconstructed Model Node Coordinates

[0051]

Claims

1. A database-based method for predicting and numerically compensating deformation in additively manufactured parts, characterized in that: (1) Establish the geometric model of the part and mesh the geometric model of the part. The total number of model elements is H. (2) Obtain the node number N based on the discrete unit and the discrete precision. i ,i=1,…,n and node coordinates D={…,(x i y i , z i ),…},i=1,…,n, where (x i y i , z i () represents the coordinates of node i; (3) Establishment and updating of the part deformation database; ① Randomly generate m printing paths, and initialize the k=1 (k≤m)th path; ② Calculate the node strain: Based on the thermo-coupling theory, simulate and analyze the additive manufacturing process of the k-th printing path to obtain the strain ε={ε1,ε2,…,ε i , …, ε n }, where ε i The total strain at node i is the simulation result, which includes numerical simulations of the temperature field and the stress field. e i =A -1 s i +αΔT i Where the total strain ε at node i i =[ε ix ε iy ε iz ε ixy ε ixy ε ixz ] T The total stress σ at node i i =[σ ix σ iy σ iz σ ixy σ ixy σ ixz ] T The coefficient of thermal expansion α = [α x α y α z 0 0 0] T E, ν, and G are Young's modulus, Poisson's ratio, and shear modulus, respectively. ③ Calculate the nodal deformation: The nodal deformation S is derived from the nodal strain in step ②. k ={s k 1,s k 2, ..., s k i , ..., s k n }, where s k i Let s be the deformation of node i along path k. i The definition is as follows: s k i =(s k xi ,s k yi ,s k zi ) s k xi s k yi s k zi Let i be the deformation of node i along path k in the x, y, and z directions, where i = 1, ..., n; ④ Update the database: Update the deformation of the node corresponding to path k to the database; ⑤ Determine whether k is less than or equal to m. If so, let k = k + 1 and repeat steps ②-④. If not, proceed to step (4). (4) Calculation of deformation weight coefficients for path nodes to be printed: Initialize matching weight coefficient w k =0, sequentially check if the T-th step (T=1,…,H-1) of the path to be printed is equal to the T-th step of each path in the database. If they are equal, then let w = 0. k =w k +1; (5) Prediction of deformation of path nodes to be printed: Predicted value s of deformation of path nodes to be printed * i Let ,i=1,…,n be the ratio of the sum of the products of the deformation and matching weight coefficients of the node under all paths in the database to the sum of all weight coefficients, where s is the predicted value of the deformation of the path node to be printed. * i The definition is as follows: (6) Compensation for the same-axis reverse deformation of the original geometric model of the part: Update the coordinates of each node, multiply the deformation of each node of the path to be printed in step (5) by -1 in the reverse direction and add it to the node coordinates in step (2) to obtain the new coordinates of each node, i.e., the reconstructed node coordinates D'. Sweep the reconstructed node coordinates to obtain the reconstructed model, where the reconstructed node coordinates D' are defined as: D'=D-S * That is: D' = {…, (x i , y i , z i ), …} - {…, (s * xi , s * yi , s * zi ), …} ={…,(x i -s * xi ,y i -s * yi ,z i -s * zi ),…}。

Images