Patent Application
20240102798
The present invention relates to a deformation prediction method for an additively manufactured object.
In recent years, there is an increasing need for shaping using a 3D printer as means for production, and research and development are proceeding toward practical use of shaping using a metal material. A 3D printer that builds a metal material uses a heat source such as laser beams, electron beams, and arcs to melt a metal powder or a metal wire, and deposits the molten metal to create an additively manufactured object.
In the production of such an additively manufactured object, a technique is known that analyzes thermal deformation that occurs during shaping by computer simulation. A result of the analysis is used to create a deposition plan that efficiently builds a high-quality additively manufactured object.
For example, Patent Literature 1 discloses a technique for analyzing a deformation amount and residual stress after welding for a welded structure having a plurality of welded layers. Patent Literature 2 discloses a technique of creating shaping data in anticipation of deformation from shape data of a structure manufactured by layered shaping. Patent Literature 3 discloses a technique for calculating an inherent strain of a structure that is built by additive manufacturing with a low calculation load. Patent Literature 4 discloses a technique for reducing a calculation time when a computer analyzes residual stress and deformation occurring in an additively manufactured object. Patent Literature 5 discloses a technique for accurately evaluating the amount of thermal deformation of an additively manufactured object in a short period of time.
All of the above-described Patent Literatures 1 to 5 employ an inherent strain method using elastic analysis that can be analyzed in a relatively short time than thermoelastic-plastic analysis.
However, it is not possible to predict and correct the thermal deformation of an object to be built that occurs in a additive manufacturing process with high efficiency prior to shaping, using the above-described analysis method. When using elasto-plastic analysis that reproduces actual phenomena such as solidification and cooling, a calculation load is large, and thus it is difficult to perform calculations in a realistic time. Even when the inherent strain method using the elastic analysis described above is adopted, the calculation time may become long when the number of shaping paths is large.
In the method, deformation of a weld bead of a next layer is sequentially calculated considering deformation of a weld bead of a previous layer. Therefore, it is necessary to proceed calculation of the deformation of the weld beads from a first layer surface to the n-th layer sequentially. Therefore, it is inevitable that the calculation time increases as the number of layers increases.
Accordingly, it is an object of the present invention to provide a deformation prediction method for an additively manufactured object, which enables analysis of deformation occurring during shaping in a short period of time, even when a shape of an additively manufactured object to be produced is complicated. Therefore, it is possible to create a deposition plan that can obtain a highly efficient and high quality additively manufactured object.
The present invention consists of the following configuration.
A deformation prediction method for an additively manufactured object that is built by repeatedly depositing a weld bead layer of a next layer on a weld bead layer formed by a weld bead obtained by melting and solidifying a filler material, the deformation prediction method including the steps of:
According to the invention, even when a shape of an additively manufactured object to manufacture is complicated, analysis of deformation which occurs at a time of shaping can be performed in a short time.
Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
The present invention predicts deformation that occurs in an additively manufactured object when an additively manufactured object is built by repeatedly depositing a next layer of a welded bead layer on a welded bead layer formed by a welded bead obtained by melting and solidifying a filler material.
Additively manufactured object Manufacturing Apparatus
First, a manufacturing procedure of an additively manufactured object will be described.
There are various manufacturing methods for manufacturing an additively manufactured object, but here, a method of depositing weld beads by arc welding will be described. The additively manufactured object manufacturing apparatus 10 is an apparatus for forming an additively manufactured object or an additively manufactured object as a raw material for obtaining a built product with a desired shape. The additively manufactured object manufacturing apparatus 10 includes an additive manufacturing device 11, a power supply device 13, and a controller 15 that controls the additive manufacturing device 11 and the power supply device 13 in an integrated manner.
The additive manufacturing device 11 includes a welding robot 19 provided with a torch 17 on a tip axis, and a filler material supply unit 23 that supplies a filler material (welding wire) M to the torch 17. The torch 17 holds the filler material M in a state of protruding from a tip.
The welding robot 19 is an articulated robot, and the torch 17 is supported so that the filler material M can be continuously supplied. Position and posture of the torch 17 can be arbitrarily set three-dimensionally within a range of degrees of freedom of a robot arm.
The torch 17 has a shield nozzle (not illustrated), and shielding gas is supplied from the shield nozzle. The arc welding method may be a consumable electrode type such as coated arc welding or carbon dioxide gas arc welding, or a non-consumable electrode type such as TIG welding or plasma arc welding, and is appropriately selected according to an additively manufactured object to be manufactured.
For example, in the case of the consumable electrode type, a contact tip is disposed inside the shield nozzle, and the contact tip holds the filler material M to which melting current is supplied. The torch 17 holds the filler material M and generates an arc from a tip of the filler material M in a shielding gas atmosphere. The filler material M is fed from the filler material supply unit 23 to the torch 17 by a feeding mechanism (not illustrated) attached to the robot arm or the like. When the torch 17 is moved and the continuously fed filler material M is melted and solidified, a linear weld bead B, which is a melted and solidified body of the filler material M, is formed on a base plate 25.
Any commercially available welding wire can be used as the filler material M. For example, a wire specified by a MAG welding and MIG welding solid wire (JIS Z 3312) for mild steel, high-strength steel, and low-temperature steel, and an arc welding flux-cored wire (JIS Z 3313) for mild steel, high-strength steel, and low-temperature steel can be used.
A heat source for melting the filler material M is not limited to the arc described above. For example, a heat source using other methods such as a heating method using both an arc and a laser beam, a heating method using plasma, a heating method using an electron beam or a laser beam, or the like may be employed. When heating with an electron beam or a laser beam, an amount of heating can be more finely controlled, so the state of the weld bead can be maintained more appropriately, and thus a quality of the additively manufactured object can be further improved.
The controller 15 includes a CAD/CAM unit 31, a trajectory calculation unit 33, a storage unit 35, and a control unit 37 to which the units are connected. The controller 15 is configured of a computer device including a CPU, a memory, a storage, and the like.
The CAD/CAM unit 31 reads three-dimensional shape data (CAD data, or the like) of the additively manufactured object to be manufactured, divides a three-dimensional model according to the three-dimensional shape data into a plurality of blocks, and generates block shape data representing a shape of each block. The trajectory calculation unit 33 divides the generated block shape data into weld bead shapes, and determines a movement trajectory of the torch 17 along each divided weld bead shape. Then, according to the block shape data, the welding conditions, and the movement trajectory of the torch 17, a drive program is created to drive each part such as the welding robot 19 of the additive manufacturing device 11, and the power supply device 13.
The storage unit 35 stores various data including information on generated block shape data, welding conditions, movement trajectory of the torch 17, and the like, and drive programs.
The control unit 37 executes the drive program stored in the storage unit 35 to drive each part of the additive manufacturing device 11. That is, the welding robot 19 drives the power supply device 13 according to a command from the control unit 37, moves the torch 17 along a trajectory set in the drive program, and generates an arc at a tip of the torch 17 at a desired timing.
The additively manufactured object manufacturing apparatus 10 having the configuration described above drives each part including the welding robot 19 and the power supply device 13 according to the created drive program, thereby forming a weld bead along the set movement trajectory of the torch 17. That is, the torch 17 is moved and the filler material M is melted, and then the melted filler material M is supplied onto the base plate 25. As a result, a bead layer in which a plurality of linear beads are solidified and arranged on the base plate 25 is formed. By depositing a similar bead layer multiple times on the bead layer, the additively manufactured object W having a multilayer structure as illustrated in
The driving program may be generated by another computer device by inputting required information to another computer device other than the controller 15. Here, the generated drive program is input to the storage unit 35 of the controller 15 via appropriate communication means such as LAN.
In a deposition process of the above-described additively manufactured object W, by analytically determining thermal deformation during weld bead formation in advance, it is possible to create a deposition plan that takes the thermal deformation into account, and to achieve higher-precision shaping.
To analyze thermal deformation of the additively manufactured object, a formation procedure of the weld bead is determined based on the shape data of the additively manufactured object and the welding conditions by the additive manufacturing device 11 described above. Then, deformation of each weld bead when deposition manufacturing is performed according to the formation procedure is analytically determined, and a deposition plan that sets various conditions such as weld bead size, bead formation path, welding speed, and welding current is created to obtain a desired target shape, and then the aforementioned drive program is created based on this deposition plan. The created drive program is stored in the storage unit 35 of the controller 15, and the control unit 37 executes the drive program to build an additively manufactured object.
Deformation Prediction Method for Additively manufactured object
Next, a deformation prediction method for a deposited product, which analytically calculates thermal deformation during shaping of the additively manufactured object described above, will be described.
Configuration of Analysis Device
An analysis device 100 is a computer device that includes a CPU 41 as a processor, a memory 43 such as random access memory (RAM) and read only memory (ROM), a storage unit 45 such as hard disk drive (HDD) and solid state drive (SSD), an input unit 47, an output unit 49, and a communication unit 51. The analysis device 100 is connected to a network 53 via the communication unit 51 and is capable of transmitting and receiving information from a server 55 or the like connected to the network 53. The controller 15 of the additive manufacturing device 11 may be connected to the network 53 so that information such as a drive program can be input/output to the additive manufacturing device 11.
A multi-core CPU capable of parallel processing is preferably used as the CPU 41, which is a processor. It is more preferable to use a simultaneous multithreading technique that treats one CPU core as a plurality of pseudo cores. Accordingly, processing can be efficiently executed by allocating processing to CPU cores recognized by an operating system (OS) or an application.
The storage unit 45 stores an inherent strain database DB (hereinafter referred to as database DB) required for the analysis described below, and a program that causes the analysis device 100 to function as a device for analyzing deformation.
The database DB has information on various conditions such as welding conditions, material properties of deposited metal, and shape data of additively manufactured objects, and information on inherent strain obtained according to measured deformation values and analysis values.
The input unit 47 may be an input device such as a keyboard and a mouse, or an interface that receives information from the outside.
The output unit 49 may be an output device such as a monitor that displays analysis results by the analysis device 100 on a screen, or an interface that outputs the results to the outside as an output signal.
The analysis device 100 includes an inherent strain DB creation/storage unit 61, a block division unit 63 for a target shape, an inherent strain definition unit 65 that determines inherent strain for each block, a parallel calculation unit 67, and a calculation result integration unit 69.
The analysis device 100 predicts deformation of an additively manufactured object that is built by repeatedly depositing the next weld bead layer on the weld bead layer formed by the weld beads made by melting and solidifying the filler materials.
Deformation Prediction Procedure Here, each process of a deformation prediction method for an additively manufactured object is described.
As illustrated in
The input shape data is sent to the block dividing unit 63, which divides the shape of the additively manufactured object into a plurality of blocks (S12). Here, an example in which the blocks are divided in units of weld bead will be described, but the present invention is not limited thereto.
Next, the inherent strain definition unit 65 determines the inherent strain for each divided block (S13). The inherent strain of each block is determined by referring to the database DB created by the inherent strain DB creation/storage unit 61.
Inherent Strain Database
Here, the inherent strain is calculated in units of paths (torch trajectories) in which the torch 17 of the additive manufacturing device 11 illustrated in
When forming the weld beads B as shown by a planned line indicated by the solid line in
When the thermal contraction is considered in relation to deformation of the built product, as illustrated in
Therefore, a matrix [H] that relates inherent strain and elastic strain (displacement) is calculated in advance according to various bead shapes, welding conditions, and the like, so that the bead shape can be analytically obtained based on the inherent strain method. A relationship between an inherent strain εK, an elastic strain εj, and a matrix H is given by Equation (1).
[Equation 1]
{εK}=[H]{εj} (1)
Specifically, as illustrated in
A relative displacement {Um}=[Δu1, Δu2, . . . , Δun]T between the path K and a path K−1 is calculated by actually depositing and measuring the weld beads (S23).
Using the relative displacement {Um} calculated by actual measurement and the matrix [H] calculated analytically, the inherent strain {ε} is calculated from Equation (2) (S24).
[Equation 2]
[H]={Um}{ε} (2)
The inherent strain {ε} can be obtained with high reliability by calculating with the least square method using a plurality of data sets under different conditions. Equation (3) is a basic equation for calculating the inherent strain {ε} by the method of least squares.
The inherent strain calculated from the above-described basic equation is registered in the database DB in association with the bead shape of the path K, welding conditions, and the like (S25). It is determined whether the path K is a final path (S26), and when there is another path, K is incremented (S27), and the inherent strain for the next path is calculated in the same manner as described above. By repeating the process up to the final path, the inherent strain for each path is registered in the database DB. The database DB is constructed by calculating analysis results and actual measurement results in various different paths and welding conditions in addition to the above-described paths, and accumulating information on the inherent strain in each case.
Returning to
The inherent strain definition unit 65 refers to the database DB created by the inherent strain DB creation/storage unit 61 to determine the inherent strain corresponding to each divided block.
Calculation of Deformation by Inherent Strain Method
When the inherent strain is determined for each block, the parallel calculation unit 67 calculates a deformation amount (deformation vector) based on the inherent strain method for each block, that is, for each path (S14). The calculation is performed by parallel processing of multiple threads, that is, simultaneous arithmetic processing of a multi-core CPU.
That is, processing for each layer, processing for calculating a deformation amount {d1} for a path 1 of a first layer, processing for calculating a deformation amount {Δd2} for a path 2 of a second layer, and processing for calculating a deformation amount {Δdn} for a path n of an n-th layer, is performed simultaneously. Calculation of the deformation amount in each path is absolute calculation of the deformation amount for the path 1 because a lower layer is the base plate 25, and for the other paths, calculation of the deformation amount is calculation relative to the lower layer. The deformation amount calculated here is a vector amount indicating a direction of deformation and a magnitude of deformation in that direction.
Integration of Deformation of Each Block
Next, the calculation results of the deformation amount in each path calculated by the simultaneous arithmetic processing are integrated (S15). That is, the deformation amount and deformation direction of the entire additively manufactured object are calculated by adding the deformation amount of each path (block) according to the deformation direction of the path. Specifically, the deformation amount {Δd2}, . . . , {Δdn} in a deposition direction of the other layers is vector-added to the deformation amount {d1} of the path of the first layer. For example, when integrating the deformation amounts of weld beads in the deposition direction, d1+Δd2+Δd3, which is the deformation amount in the deposition direction, is a height of three layers of beads, as illustrated in
Then, the calculated deformation amount is output as the predicted deformation amount of the additively manufactured object (S16). The predicted deformation amount may be output from the output unit 49 illustrated in
According to the deformation prediction method, since the deformation is predicted using the inherent strain method, calculation of elasto-plastic analysis, which makes the calculation complicated, is not performed, so that a calculation load can be reduced.
Other Division Examples
A process of dividing the shape of the additively manufactured object described above into a plurality of blocks is an example in which a weld bead is divided as a unit and a shape of one block group in which each block is combined becomes the shape of an additively manufactured object, but the process is not limited thereto. For example, a deposit body of a plurality of weld beads may be divided into blocks as a unit. Here, the shape of the additively manufactured object may consist of a plurality of block groups.
The additively manufactured object Wa includes a frame portion 71 built by depositing weld beads B1 on the base plate 25 and an internal built portion 73 built inside the frame portion 71 by weld beads B2. The internal built portion 73 is configured by depositing weld bead layers made of the weld beads B2.
Such an additively manufactured object Wa is formed, for example, by disposing two different torches for the torch 17 illustrated in
Instead of changing the type of the filler material that forms the weld bead, welding conditions such as the welding speed, the welding current, and the welding voltage may be changed, or it is also possible to change both the type of the filler material and the welding conditions. In either case, weld beads having different characteristics are formed.
Thus, when the additively manufactured object has the weld beads B1 and B2 with different characteristics, the frame portion 71 built from the weld beads B1 and the internal built portion 73 built from the weld beads B2 are divided into separate blocks. Then, as described above, the deformation amount and deformation direction of each block before and after weld bead formation are calculated by parallel processing using a plurality of processors based on the inherent strain method for each block.
Accordingly, since the deformation is calculated with the frame portion 71 and the internal built portion 73 as separate blocks, the weld beads having similar characteristics become the same block, and thus the amount of calculation can be further reduced. When leaving the frame portion 71 as it is and forming a weld bead with the internal built portion 73 under different conditions, deformation of the entire additively manufactured object can be easily calculated simply by adding the newly calculated deformation of the internal built portion 73 to the deformation of the block of the frame portion 71 that has already been calculated.
In the above example, blocks are divided according to the difference in weld bead characteristics, but blocks can also be divided according to the shape of the weld bead. For example, a deformation pattern of each block can be simplified by calculating the deformation of a weld bead formed along a straight line and a weld bead formed along a curve as separate blocks. It becomes possible to easily analyze the inherent strain, thereby reducing the computational load for the analysis.
The deformation of the additively manufactured object was predicted using the verification model MDL illustrated in
1. Verification Details
(1) Comparison between program corresponding to present invention and general-purpose software
When a program based on the method for predicting deformation of an additively manufactured object according to the present invention is executed to perform parallel calculation for each block (Test Example 1), Computation times were compared between the case (Test Example 1) of executing parallel calculation for each block by executing a program based on the deformation prediction method for an additively manufactured object according to the present invention and the case (Test Example 2) of executing parallel calculation automatically by the CPU itself using predetermined general-purpose software. However, the CPU configuration of the analysis device used is set such that a CPU using the general-purpose software can perform faster processing than a CPU using a program based on the deformation prediction method for an additively manufactured object according to the present invention.
(2) Presence of Parallel Calculation
In Test Example 1 described above, the performance of the CPU of the analysis device was set to be the same, and the calculation times were compared between the case (Test Example 1-1) without parallel calculation for each block and the case (Test Examples 1-2, 1-3, 1-4) with parallel calculation for each block.
2. Used Hardware
(1) Specification of analysis device [Configuration 1] that executes program corresponding to present invention
CPU: Intel Core (registered trademark) i7-6800K (6 cores, 12 threads, operating clock 3.4 GHz)
OS: Microsoft Windows 10 (registered trademark)
(2) Specification of analysis device [Configuration 2] that executes general-purpose software
CPU: Intel Xeon (registered trademark) E5-2637v4 (8 cores, 16 threads, operating clock 3.50 GHz)
OS: SUSE Linux (registered trademark) Enterprise Server 11 SP4
3. Calculation Model
3D model of deposited built body, which is wall object composed of 20 layers of weld beads (number of nodes: 79940, number of elements: 63954)
4. Verification Result
In Test Examples 1-1, 1-2, 1-3, and 1-4, the analysis device of Configuration 1 was used. The calculation time was 8 minutes and 27 seconds when parallel calculation was not performed for each block in Test Example 1-1 (calculated by one thread). The calculation time when the parallel calculation for each block in Test Example 1-2 was performed with two threads was 4 minutes and 18 seconds, the calculation time was 2 minutes and 6 seconds when 6 threads were used in Test Example 1-3, and the calculation time was 1 minute and 35 seconds when 12 threads were used in Test Example 1-4.
On the other hand, in Test Example 2 using the analysis device of Configuration 2, the automatic parallel calculation of the CPU was performed using the predetermined general-purpose software. As a result, the calculation time was 2 minutes and 48 seconds, despite the configuration having higher arithmetic processing performance than the analysis device of Configuration 1.
From the above results, by simultaneously calculating the deformation of a plurality of blocks (weld beads), the calculation time can be shortened as the number of parallel threads increases. Compared to the case (Test Example 2) of automatic parallel calculation of the CPU using the general-purpose software, the calculation time of the parallel calculation with 12 threads in Test Example 1-4 was shorter. This is assumed to be due to the difference in parallel calculation algorithms between the two cases, that is, the general-purpose software computes sequential deformation for each weld bead, while the present method simultaneously computes for a plurality of weld beads.
Thus, the present invention is not limited to the above-described embodiment, and combinations of the configurations of the embodiment with each other, and modifications and applications by those skilled in the art based on the descriptions in the specification and well-known techniques are also contemplated by the present invention and are included in the scope of protection.
As described above, the following matters are disclosed in the specification.
(1) A deformation prediction method for an additively manufactured object that is built by repeatedly depositing a weld bead layer of a next layer on a weld bead layer formed by a weld bead obtained by melting and solidifying a filler material, the deformation prediction method including the steps of:
According to the deformation prediction method for the additively manufactured object, simultaneously calculating the deformation amount and deformation direction for each of the plurality of blocks by parallel calculation enables high-speed arithmetic processing, and thus deformation prediction can be performed in a short time even for a complicated shape of the additively manufactured object. Calculation load can be reduced because the inherent strain method is used without the need for calculation of elasto-plastic analysis, which makes calculations complicated.
(2) The deformation prediction method for the additively manufactured object according to (1), where
According to the deformation prediction method for the additively manufactured object, since deformation is predicted by dividing blocks for each weld bead, finer deformation can be accurately predicted.
(3) The deformation prediction method for the additively manufactured object according to (1), where
According to the deformation prediction method for the additively manufactured object, each deposit body of the weld beads is divided into blocks to predict deformation. Therefore, for example, in the case of an additively manufactured object in which only specific blocks are replaced, the deformation of the entire additively manufactured object can be easily calculated by adding the deformation of the replaced blocks to the deformation of the other blocks.
(4) The deformation prediction method for the additively manufactured object according to any one of (1) to (3), where
According to the deformation prediction method for the additively manufactured object, a deformation pattern of each block can be simplified, and thus the analysis of the inherent strain can be simplified.
(5) The deformation prediction method for the additively manufactured object according to any one of (1) to (4), where
According to the deformation prediction method for the additively manufactured object, even when the additively manufactured object has a complicated shape, the deformation is added for each block group, so that complexity of calculation of the deformation of the additively manufactured object can be reduced.
The application is based on a Japanese patent application (Japanese Patent Application No. 2021-13576) filed on Jan. 29, 2021, the content of which is incorporated herein by reference.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2021-013576 | Jan 2021 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2022/000398 | 1/7/2022 | WO |
