Deformation prediction method for layered structures

Abstract

The deformation prediction method for stacked structures includes the steps of dividing the shape of the stacked structure into multiple blocks, calculating the deformation amount and deformation direction of each block before and after the formation of the weld bead based on the inherent strain method through parallel processing of multiple threads, setting at least one block group composed of interlocking blocks from the multiple blocks, and calculating the overall deformation of the block group by adding the deformation amounts of each block constituting the block group according to the deformation direction of the block.

Description

Technical Field

[0001] This invention relates to a method for predicting the deformation of stacked structures. Background Technology

[0002] In recent years, the demand for 3D printers as a means of production has increased, and research and development are advancing towards the practical application of 3D modeling using metal materials. 3D printers that model metal materials use heat sources such as lasers, electron beams, and electric arcs to melt metal powder or filaments, and then layer the molten metal to create stacked models.

[0003] There are known techniques for using computer simulation to analyze thermal deformation that occurs during the creation of such layered structures. The results of this analysis are used to develop layering plans that efficiently create high-quality, layered structures.

[0004] For example, Patent Document 1 discloses a technique for analyzing the deformation and residual stress of a welded structure having multiple weld layers after welding. Patent Document 2 discloses a technique for generating shape data by predicting deformation based on shape data of a structure manufactured through layered molding.

[0005] Patent document 3 discloses a technique for calculating the inherent strain of a structure shaped by layering with low computational load. Patent document 4 discloses a technique for reducing the computation time for analyzing residual stress and deformation generated in layered structures using computers. Patent document 5 discloses a technique for accurately evaluating the amount of thermal deformation of layered structures in a short time.

[0006] The aforementioned patent documents 1 to 5 all employ the intrinsic strain method of elastic analysis, which can be performed in a shorter time compared to thermo-elastic-plastic analysis.

[0007] Existing technical documents

[0008] Patent documents

[0009] Patent Document 1: Japanese Patent Application Publication No. 2014-115789

[0010] Patent Document 2: Japanese Patent Application Publication No. 2020-027491

[0011] Patent Document 3: Japanese Patent Application Publication No. 2018-184623

[0012] Patent Document 4: International Publication No. 2019 / 049981

[0013] Patent Document 5: International Publication No. 2018 / 123858 Summary of the Invention

[0014] The problem that the invention aims to solve

[0015] However, the analytical methods described above cannot efficiently predict and correct for thermal deformation of the object being molded during the layering molding process before molding. Furthermore, using elasto-plastic analysis, which reproduces actual phenomena such as solidification and cooling, results in a large computational load that is difficult to achieve within a realistic timeframe. Moreover, even when employing the inherent strain method using elastic analysis, the computation time can sometimes increase with a large number of molding passes.

[0016] exist Figure 10 The diagram illustrates the analytical sequence for thermal deformation of a laminated structure using the conventional method of inherent strain. When forming a laminated structure by stacking multiple weld beads, firstly, the inherent strain of the first weld bead, which forms the initial layer, is determined by referring to a pre-prepared database of inherent strains. Then, the deformation {d1} of the first weld bead is calculated using the determined inherent strain. Next, the deformation {d2} of the second weld bead is calculated based on the inherent strain of the second weld bead obtained from the database and the deformation {d1} of the first weld bead. This process is repeated until the deformation {dn} of the nth weld bead, which forms the final layer, is calculated, thereby determining the overall deformation of the laminated structure.

[0017] In this method, the deformation of the previous layer's weld beads is considered, and the deformation of the next layer's weld beads is calculated sequentially. This requires progressively advancing the calculation of the deformation of weld beads from the first layer to the nth layer. Therefore, an increase in calculation time with the increase in the number of layers is unavoidable.

[0018] Therefore, the object of the present invention is to provide a deformation prediction method for layered shapes that can analyze the deformations generated during the shaping process in a short time, even if the shape of the layered shape is complex, and to produce layering plans that can efficiently obtain high-quality layered shapes.

[0019] Solution for solving the problem

[0020] The present invention is composed of the following structure.

[0021] A method for predicting the deformation of a layered structure is provided, which is a method for predicting the deformation of a layered structure formed by repeatedly stacking layers of weld beads (formed by melting and solidifying filler material) onto the next layer of weld beads.

[0022] The deformation prediction method for the layered structure includes:

[0023] The process of dividing the shape of the stacked structure into multiple blocks;

[0024] The process of calculating the deformation amount and deformation direction of each block before and after the formation of the weld bead based on the inherent strain method through parallel processing of multiple threads;

[0025] The process of setting up at least one block group consisting of interlocking blocks from the plurality of blocks; and

[0026] The process of calculating the overall deformation of the block group by adding the deformation amounts of each block that makes up the block group according to the deformation direction of that block.

[0027] Invention Effects

[0028] According to the present invention, even if the shape of the layered object is complex, the deformation generated during the shaping process can be analyzed in a short time. Attached Figure Description

[0029] Figure 1 This is a schematic diagram showing the manufacturing apparatus for layered shapes.

[0030] Figure 2 This is a hardware structure diagram of the analytical device for deformation prediction.

[0031] Figure 3 It is shown Figure 2 The diagram shows the functional structure of the analytical device.

[0032] Figure 4 This is a flowchart illustrating the process of predicting the deformation of a stacked structure.

[0033] Figure 5 This is a flowchart illustrating the production sequence of the database of inherent strain.

[0034] Figure 6 The diagrams (A) and (B) schematically illustrate the strain in the case of stacked weld beads.

[0035] Figure 7 It is a schematic cross-sectional view of a stacked structure created using different types of weld beads.

[0036] Figure 8 This is an explanatory diagram showing a verification model used to predict the deformation of stacked structures.

[0037] Figure 9 It is a graph showing the calculation time for test examples 1-1, 1-2, 1-3, 1-4 and test example 2.

[0038] Figure 10 This is an explanatory diagram showing the analytical sequence of thermal deformation of layered structures using the inherent strain method. Detailed Implementation

[0039] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0040] The present invention predicts the deformation that will occur in the stacked object when a layer of weld beads formed by melting and solidifying filler material is repeatedly stacked onto another layer of weld beads to form a stacked object.

[0041]

[0042] First, the manufacturing sequence of the layered structure will be explained.

[0043] Figure 1 This is a schematic diagram showing the manufacturing apparatus for layered shapes.

[0044] Various manufacturing methods exist for the production of laminated shapes, but here we will describe the method of depositing weld beads by arc welding. The laminated shape manufacturing apparatus 10 is an apparatus for forming laminated shapes or as a rough material for obtaining a shape of a desired shape, and includes a laminated shape apparatus 11, a power supply apparatus 13, and a controller 15 for comprehensively controlling the laminated shape apparatus 11 and the power supply apparatus 13.

[0045] The stacking molding device 11 includes a welding robot 19 with a welding torch 17 on its front end shaft and a filler material supply unit 23 that supplies filler material (welding wire) M to the welding torch 17. The welding torch 17 holds the filler material M in a state that protrudes from the front end.

[0046] The welding robot 19 is a multi-jointed robot that supports the filler material M in a manner that allows for continuous supply within the welding torch 17. The position and orientation of the welding torch 17 can be arbitrarily set in three dimensions within the range of degrees of freedom of the robotic arm.

[0047] The welding torch 17 has a conductive nozzle (not shown) and supplies shielding gas from the conductive nozzle. As an arc welding method, it can be any of the consumable electrode types such as covered arc welding or carbon dioxide gas arc welding, or the non-consumable electrode types such as TIG welding or plasma arc welding, and can be appropriately selected according to the stacked shape being produced.

[0048] For example, in the case of a consumable electrode type, a conductive tip is disposed inside the conductive nozzle, and the filler material M supplied with molten current is held in the conductive tip. While holding the filler material M, the welding torch 17 generates an arc from the front end of the filler material M under a shielding gas atmosphere. The filler material M is fed from the filler material supply section 23 to the welding torch 17 via a delivery mechanism (not shown) mounted on a robotic arm or the like. Furthermore, as the continuously fed filler material M melts and solidifies while the welding torch 17 is moved, a linear weld bead B, which is the molten solidified filler material M, is formed on the base plate 25.

[0049] As filler material M, all commercially available welding wires can be used. For example, welding wires specified such as solid welding wires for MAG welding and MIG welding of mild steel, high-tensile steel and low-temperature steel (JIS Z3312) and flux-cored welding wires for arc welding of mild steel, high-tensile steel and low-temperature steel (JIS Z 3313) can be used.

[0050] It should be noted that the heat source for melting the filler material M is not limited to the electric arc described above. For example, other heat sources based on methods such as heating with both electric arc and laser, heating with plasma, or heating with electron beam or laser can also be used. When heating with electron beam or laser, the amount of heat can be controlled more precisely, and the state of the weld bead can be maintained more appropriately, which helps to further improve the quality of the laminated structure.

[0051] The controller 15 includes a CAD / CAM unit 31, a track calculation unit 33, a storage unit 35, and a control unit 37 connected to them. The controller 15 is composed of a computer device equipped with a CPU, memory, storage devices, etc.

[0052] The CAD / CAM unit 31 reads in the three-dimensional shape data (CAD data, etc.) of the stacked model to be created, and divides the solid model corresponding to the three-dimensional shape data into multiple blocks, generating block shape data representing the shape of each block. The trajectory calculation unit 33 divides the generated block shape data into weld bead shapes and determines the movement trajectory of the welding torch 17 along the shapes of each weld bead. Furthermore, based on the block shape data, welding conditions, and the movement trajectory of the welding torch 17, a driver program is created to drive the welding robot 19, power supply unit 13, and other components of the stacked modeling device 11.

[0053] The storage unit 35 stores various data and drivers, including information such as the generated block shape data, welding conditions, and the movement trajectory of the welding torch 17.

[0054] The control unit 37 executes the driver program stored in the storage unit 35, thereby driving each part of the stacked modeling device 11. That is, the welding robot 19 drives the power supply unit 13 according to the instructions from the control unit 37, and moves the welding torch 17 along the track set by the driver program, and generates an electric arc at the tip of the welding torch 17 at the desired time.

[0055] The manufacturing apparatus 10 for the aforementioned layered structure drives the components, including the welding robot 19 and the power supply unit 13, according to a pre-programmed drive program, thereby forming weld beads along the predetermined trajectory of the welding torch 17. Specifically, the welding torch 17 is moved, and the filler material M is melted, which is then supplied onto the base plate 25. This forms a weld bead layer on the base plate 25, consisting of multiple solidified and arranged linear weld beads. The same weld bead layer is then repeatedly stacked on top of this weld bead layer to create the desired shape. Figure 1 The multi-layered structure W shown is a multi-layered object.

[0056] It should be noted that the driver can also input the required information to other computer devices different from the controller 15, and be generated by those other computer devices. In this case, the generated driver is input to the storage unit 35 of the controller 15 via a suitable communication means such as a LAN.

[0057] In the aforementioned stacking process of the stacked model W, by pre-analytically calculating the thermal deformation during the formation of the weld bead, it is possible to create a stacking plan that takes into account this thermal deformation and achieve a higher precision model.

[0058] To analyze the thermal deformation of the stacked model, the formation sequence of the weld beads is determined based on the shape data of the stacked model and the welding conditions based on the stacking molding device 11. Furthermore, the deformation of each weld bead in the case of stacking in this formation sequence is analytically calculated. To obtain the desired target shape, a stacking plan is created, setting conditions such as weld bead size, number of weld bead formation passes, welding speed, and welding current. Based on this stacking plan, the aforementioned driver program is created. The created driver program is stored in the storage unit 35 of the controller 15, and the driver program is executed by the control unit 37 to shape the stacked model.

[0059] <Methods for Predicting Deformation of Layered Structures>

[0060] Next, a method for predicting the deformation of stacked objects by analyzing the thermal deformation of the aforementioned stacked objects will be explained.

[0061] (Structure of the analytical apparatus)

[0062] Figure 2 This is a hardware structure diagram of the analytical device for deformation prediction.

[0063] The parsing device 100 is a computer device comprising a CPU 41 as a processor, memory 43 such as RAM (Random Access Memory) and ROM (Read Only Memory), storage units 45 such as HDD (Hard Disk Drive) and SSD (Solid State Drive), an input unit 47, an output unit 49, and a communication unit 51. The parsing device 100 is connected to a network 53 via the communication unit 51 and can send and receive information from a server 55 or the like connected to the network 53. Furthermore, the controller 15 of the aforementioned modeling device 11 can be connected to the network 53, enabling the input / output of driver programs and other information to the modeling device 11.

[0064] The CPU41, as the processor, preferably uses a multi-core CPU capable of parallel processing. Furthermore, it is more preferable to use a multi-threaded technique that virtually treats one CPU core as multiple cores. According to this technique, processing is allocated to the CPU cores identified by the OS (Operating System) or the application, enabling efficient processing.

[0065] The storage unit 45 stores the inherent strain database DB (hereinafter referred to as database DB) required for analysis, as described later, and the program that enables the analysis device 100 to function as a device for performing deformation analysis.

[0066] The database DB contains information on various conditions such as welding conditions, material properties of the weld metal, shape data of the stacked structure, and inherent strain calculated based on measured and analytical values ​​of deformation.

[0067] In addition to being an input device such as a keyboard or mouse, the input unit 47 can also be an interface for receiving information from the outside.

[0068] In addition to being an output device such as a monitor that displays the resolution results based on the resolution device 100 on a screen, the output unit 49 can also be an interface that outputs signals to the outside.

[0069] Figure 3 It is shown Figure 2 The functional block diagram of the analytical device 100 shown is a functional structure diagram.

[0070] The analysis device 100 includes an inherent strain DB creation / storage unit 61, a target shape block segmentation unit 63, an inherent strain definition unit 65 that determines the inherent strain of each block, a parallel calculation unit 67, and a calculation result synthesis unit 69.

[0071] The analytical device 100 predicts the deformation of a stacked object formed by repeatedly stacking layers of weld beads, which are formed by melting and solidifying filler material, onto the next layer of weld beads.

[0072] (Deformation prediction order)

[0073] Here, we will explain the steps involved in the deformation prediction method for layered structures.

[0074] Figure 4 This is a flowchart illustrating the process of predicting the deformation of a stacked structure. In the following description, appropriate references will be made. Figure 2 , Figure 3 The structure of the analytical device 100 shown.

[0075] like Figure 4 As shown, firstly, the shape data of the stacked model to be manufactured is input to the storage unit 45 of the analysis device 100 via the input unit 47 or the communication unit 51 (S11).

[0076] The input shape data is sent to the block division unit 63, which divides the shape of the stacked model into multiple blocks (S12). This example illustrates that the blocks are divided in units of weld beads, but it is not limited to this.

[0077] Next, the inherent strain is determined according to the segmented blocks using the inherent strain definition unit 65 (S13). The inherent strain of each block is determined with reference to the database DB created by the inherent strain DB creation / storage unit 61.

[0078] (Database of Inherent Strain)

[0079] Figure 5 This is a flowchart illustrating the creation sequence of the database DB with inherent strain.

[0080] Here, so that Figure 1 The inherent strain is calculated in units of the number of passes (torch trajectory) of the welding torch 17 moving to form a weld bead in the layered molding device 11 shown.

[0081] Figure 6 This is an illustration diagram showing the strain in the case of a multilayer weld bead B, as shown in (A) and (B).

[0082] exist Figure 6 As shown by the solid lines in (A), when forming weld beads B, the first layer (K=1) weld bead B1, the second layer (K=2) weld bead B2, and the third layer (K=3) weld bead B3 change into shapes represented by dashed lines due to thermal shrinkage after formation.

[0083] When considering thermal shrinkage in terms of the deformation of the object, Figure 6 As illustrated in (B) with a cylindrical shape, the planned shape, represented by solid lines, differs from the actual shape, represented by dashed lines, regarding the height of the shape. That is, the planned shape is indicated by reference numeral 57, but due to thermal shrinkage, the shape becomes the shape indicated by reference numeral 59, represented by dashed lines. Thermal shrinkage varies depending on the weld bead shape or welding conditions of each layer, resulting in different strains in each layer.

[0084] Therefore, by analytically determining the weld bead shape based on the inherent strain method, a matrix [H] is pre-calculated to establish the relationship between inherent strain and elastic strain (displacement) according to various weld bead shapes and welding conditions. Inherent strain ε K Elastic strain ε j The relationship between matrix H is as shown in equation (1).

[0085] [Mathematical Expression 1]

[0086] {ε K}=[H]{ε j}…(1)

[0087] Specifically, such as Figure 5 As shown, for pass K (initially the pass of the initial layer with K=1), the weld bead has a unit inherent strain ε. j In such cases, for example, the degree of deformation can be analytically determined using FEM (Finite Element Method) (S21). Based on the obtained deformation and the unit natural strain ε... j (=1), construct the matrix [H](S22) described above. Here, the matrix [H] is composed of [H]=[h1, h2, …, h n ] T (1, 2, ..., n refers to the track number).

[0088] On the other hand, the relative displacement {U} between pass K and pass K-1 is determined by actual measurement of the actual layer-by-layer fusion weld bead. m}=[Δu1,Δu2,…,Δu n ] T (S23)

[0089] Using the relative displacement {U} obtained through actual measurement m The inherent strain {ε}(S24) is calculated from the analytically obtained matrix [H] and the inherent strain {ε}(S24) according to equation (2).

[0090] [Mathematical Expression 2]

[0091] [H] = {U m}{ε}…(2)

[0092] In addition, the inherent strain {ε} is obtained by using the least squares method with multiple data sets under different conditions, thus obtaining a highly reliable value. Equation (3) is the basic formula for obtaining the inherent strain {ε} using the least squares method.

[0093] [Mathematical Expression 3]

[0094]

[0095] The inherent strain calculated based on the above basic formula is associated with the weld bead shape and welding conditions of pass K and registered in the database DB (S25). It is then determined whether pass K is the final pass (S26). If more passes exist, K is incremented (S27), and the inherent strain for the next pass is calculated in the same manner as before. This process is repeated until the final pass, thereby registering the inherent strain of each pass in the database DB. By calculating analytical and experimental results for various passes and welding conditions other than those mentioned above, and accumulating information on the inherent strain under various conditions, the database DB is constructed.

[0096] return Figure 4 The explanation will proceed from the step of determining the inherent strain according to the block (S13).

[0097] The inherent strain definition unit 65 determines the inherent strain corresponding to each segmented block by referring to the database DB created by the inherent strain DB creation / storage unit 61.

[0098] (Deformation calculation based on the inherent strain method)

[0099] When the inherent strain is determined for each block, the parallel computing unit 67 calculates the deformation (deformation vector) for each block, i.e., according to the pass, based on the inherent strain method (S14). This calculation is performed through parallel processing of multiple threads, i.e., simultaneous computation processing of a multi-core CPU.

[0100] That is, the processing for calculating the deformation amount {d1} for the first layer's pass 1, the processing for calculating the deformation amount {Δd2} for the second layer's pass 2, and the processing for calculating the deformation amount {Δdn} for the nth layer's pass n are performed simultaneously for each layer. It should be noted that the calculation of the deformation amount in each pass, for pass 1, is an absolute calculation because the lower layer is the base plate 25; for other passes, it is a calculation of the deformation amount relative to the lower layer. The deformation amount calculated here is a vector representing the deformation direction and the magnitude of the deformation in that direction.

[0101] (Combined deformations of each block)

[0102] Next, the calculation results of the deformation amounts obtained from each pass through simultaneous processing are combined (S15). That is, the deformation amounts of each pass (block) are added together according to the deformation direction of that pass to calculate the overall deformation amount and deformation direction of the laminated structure. Specifically, the deformation amounts {Δd2}, ..., {Δdn} of the lamination directions of other layers are added to the vector of the deformation amount {d1} of the first layer's pass. For example, when combining the deformation amounts related to the lamination direction of the weld bead, such as... Figure 6 As shown, the deformation amount d1+Δd2+Δd3 in the stacking direction becomes the height of the three weld layers.

[0103] Furthermore, the calculated deformation is output as the predicted deformation of the stacked model (S16). The output of the predicted deformation can be obtained from... Figure 2 The output unit 49 shows numerical values ​​to a monitor or similar device, and can also display the shape in conjunction with the state before weld shrinkage. Furthermore, the predicted deformation amount is sent to the server 55 or controller 15 via the communication unit 51, allowing the server 55 or controller 15 to correct the modeling plan for manufacturing the stacked model by incorporating the predicted deformation amount. This enables the generation of a driver program capable of performing more precise stacked modeling.

[0104] According to this deformation prediction method, the inherent strain method is used to predict deformation, thus avoiding the computationally complex calculations of elastoplastic analysis, thereby reducing the computational load.

[0105] <Other Segmentation Examples>

[0106] The process of dividing the aforementioned stacked structure into multiple blocks is an example of dividing it by weld bead and combining the blocks to form a block group, but it is not limited to this. For example, it is also possible to divide the stacked body with multiple weld beads into blocks. In this case, the shape of the stacked structure may also be composed of multiple block groups.

[0107] Figure 7 It is a schematic cross-sectional view of a stacked structure created using different types of weld beads.

[0108] The stacked shape Wa has a frame portion 71 formed by stacking weld beads B1 on the base plate 25, and an internal shape portion 73 formed by weld beads B2 inside the frame portion 71. The internal shape portion 73 is formed by stacking weld beads B2.

[0109] Such layered shapes, for example, are created by... Figure 1The welding torch 17 shown is configured as two different torches, and different types of filler material M are supplied to each torch. Specifically, the filler material forming weld bead B2 contains more sulfur (S) than the filler material forming weld bead B1. As a result, weld bead B2 has a wider and flatter cross-sectional shape than weld bead B1, resulting in better fusion with the substrate. On the other hand, weld bead B1, with less sulfur (S), has higher viscosity and surface tension, resulting in a narrower and more raised cross-sectional shape.

[0110] Alternatively, one can change the type of filler material used to form the weld bead, and instead change the welding conditions such as welding speed, welding current, and welding voltage. Alternatively, both the type of filler material and the welding conditions can be changed. In any case, weld beads with different characteristics will be formed.

[0111] In this way, when the stacked structure has weld beads B1 and B2 with different characteristics, the frame portion 71 formed by weld bead B1 and the internal shape portion 73 formed by weld bead B2 are divided into different blocks. Furthermore, as described above, the deformation amount and deformation direction of each block before and after the formation of the weld bead are calculated using parallel processing of multiple processors based on the inherent strain method.

[0112] Accordingly, since the deformation is calculated by treating the frame portion 71 and the internal shaping portion 73 as different blocks, weld beads with similar characteristics become identical blocks, further reducing the computational workload. Furthermore, when the frame portion 71 is kept in its original state and the internal shaping portion 73 is formed into weld beads under different conditions, the overall deformation of the stacked object can be easily calculated simply by adding the deformation of the calculated frame portion 71 block to the newly calculated deformation of the internal shaping portion 73.

[0113] Furthermore, while the above example illustrates segmentation based on differences in the characteristics of the weld bead, segmentation based on the shape of the weld bead is also possible. For instance, by treating weld beads formed along straight lines and weld beads formed along curves as different blocks for deformation calculation, the deformation patterns of each block can be simplified. This facilitates the analysis of inherent strain and reduces the computational burden of the analysis.

[0114] Example

[0115] Figure 8 This is an explanatory diagram showing a validation model MDL used to predict the deformation of stacked structures.

[0116] use Figure 8 The validation model MDL was used to predict the deformation of the stacked structure, and its effect of reducing computation time was confirmed.

[0117] 1. Verification Content

[0118] (1) Comparison of the program corresponding to this invention with general software

[0119] The computation time was compared between the case where parallel computation of each block was performed using the program executing the deformation prediction method for stacked structures based on the present invention (Example 1) and the case where the CPU automatically performed parallel computation using established general-purpose software (Example 2). Regarding the CPU architecture of the analysis device used, compared to the CPU architecture of the analysis device using the program executing the deformation prediction method for stacked structures based on the present invention, the CPU architecture of the analysis device using general-purpose software was configured to perform higher-speed processing.

[0120] (2) The presence or absence of parallel computing

[0121] The CPU performance of the parsing device in the above-mentioned Test Example 1 was made the same, and the computation time was compared between the case without parallel computation of each block (Test Example 1-1) and the case with parallel computation of each block (Test Examples 1-2, 1-3, 1-4).

[0122] 2. Using hardware

[0123] (1) Specifications of the parsing device [structure 1] for executing the program corresponding to this invention

[0124] CPU: Intel Core i7-6800K (6 cores, 12 threads, 3.4GHz)

[0125] OS: Windows 10 (registered trademark) by Microsoft Corporation

[0126] (2) Specifications of the parsing device [structure 2] for executing general-purpose software

[0127] CPU: Intel Xeon E5-2637v4 (8 cores, 16 threads, 3.50GHz)

[0128] OS: SUSE Linux (registered trademark) Enterprise Server 11 SP4

[0129] 3. Computational Model

[0130] A 3D model of a wall structure consisting of 20 layers of weld beads (79,940 nodes, 63,954 elements).

[0131] 4. Verification Results

[0132] Figure 9 It is a graph showing the calculation time for test examples 1-1, 1-2, 1-3, 1-4 and test example 2.

[0133] In Experiments 1-1, 1-2, 1-3, and 1-4, the parsing apparatus of Structure 1 was used. The computation time for Experiment 1-1 without parallel computation of each block (computed with 1 thread) was 8 minutes and 27 seconds. The computation time for Experiment 1-2 with parallel computation of each block using 2 threads was 4 minutes and 18 seconds. The computation time for Experiment 1-3 with 6 threads was 2 minutes and 6 seconds. The computation time for Experiment 1-4 with 12 threads was 1 minute and 35 seconds.

[0134] On the other hand, in Experiment 2, which used the analytical apparatus of Structure 2, the calculation was performed using predetermined general-purpose software and the automatic parallel computation of the CPU. The result was that, despite having higher computational performance compared to the analytical apparatus of Structure 1, the computation time was 2 minutes and 48 seconds.

[0135] Based on the above results, by simultaneously calculating the deformation of multiple blocks (weld bead) the computation time can be reduced as the number of parallel threads increases. Furthermore, compared to the case of automatic parallel computation using general-purpose software (Experiment 2), the computation time of the case in Experiments 1-4, which used 12 threads for parallel computation, was shorter. This is attributed to the difference in the parallel computation algorithms used in the two methods; that is, in general-purpose software, deformation is calculated sequentially according to the weld bead, while in this method, multiple weld beads are calculated simultaneously.

[0136] Thus, the present invention is not limited to the above-described embodiments. Combining the various structures of the embodiments with each other, as well as making changes and applications based on the description and well-known techniques by those skilled in the art, are also intended by the present invention and are included within the scope of the claimed protection.

[0137] As stated above, the following matters are disclosed in this specification.

[0138] (1) A method for predicting the deformation of a layered structure, wherein the layered structure is formed by repeatedly stacking layers of weld beads formed by melting and solidifying filler material onto the next layer of weld beads, wherein...

[0139] The deformation prediction method for the layered structure includes:

[0140] The process of dividing the shape of the stacked structure into multiple blocks;

[0141] The process of calculating the deformation amount and deformation direction of each block before and after the formation of the weld bead based on the inherent strain method through parallel processing of multiple threads;

[0142] The process of setting up at least one block group consisting of interlocking blocks from the plurality of blocks; and

[0143] The process of calculating the overall deformation of the block group by adding the deformation amounts of each block that makes up the block group according to the deformation direction of that block.

[0144] According to the deformation prediction method for this layered structure, the deformation amount and direction of each block in multiple blocks are calculated simultaneously through parallel computation, enabling high-speed computational processing. Even complex layered structures can have their deformation predicted in a short time. Furthermore, the computational load is reduced by using the inherent strain method, eliminating the need for complex elastoplastic analytical calculations.

[0145] (2) According to the deformation prediction method for layered structures described in (1), wherein,

[0146] The process of dividing the material into multiple blocks divides the shape of the stacked structure in units of the weld beads.

[0147] According to the deformation prediction method of this layered structure, deformation is predicted according to the weld bead segmentation blocks, thus enabling accurate prediction of more subtle deformations.

[0148] (3) According to the deformation prediction method for layered structures described in (1), wherein,

[0149] The process of dividing into multiple blocks divides the shape of the stacked structure into units of multiple weld beads.

[0150] According to the deformation prediction method of this laminated structure, the deformation is predicted according to the segmentation blocks of the laminated body of the weld bead. Therefore, for example, in the case of a laminated structure in which only a specific block has been replaced, the overall deformation of the laminated structure can be easily calculated by adding the deformation of the replaced block to the deformation of the other blocks.

[0151] (4) The deformation prediction method for the stacked structure according to any one of (1) to (3), wherein,

[0152] Blocks that have the same welding conditions among the plurality of blocks are set to the same block group.

[0153] Based on the deformation prediction method of this layered structure, the deformation pattern of each block is simplified, making it easy to analyze the inherent strain.

[0154] (5) The deformation prediction method for the stacked structure according to any one of (1) to (4), wherein,

[0155] When the shape of the stacked structure is divided into multiple block groups.

[0156] The deformation of the stacked structure is predicted by adding the deformation amounts of the multiple block groups according to the deformation direction of the block group.

[0157] According to the deformation prediction method of this layered structure, even if the layered structure is a complex shape, the complexity of the deformation calculation of the layered structure can be reduced because the deformation is added according to the block group.

[0158] It should be noted that this application is based on Japanese patent application (Japanese Patent Application No. 2021-13576) filed on January 29, 2021, the contents of which are referenced in this application.

[0159] Explanation of reference numerals in the attached figures

[0160] Manufacturing apparatus for 10-layer stacked objects

[0161] 11-layered structural installation

[0162] 13 Power Supply Unit

[0163] 15 controllers

[0164] 17 Welding Torch

[0165] 19 welding robots

[0166] 23 Filler Material Supply Department

[0167] 25 base plate

[0168] 31 CAD / CAM Department

[0169] 33-track computing unit

[0170] 35 Storage Unit

[0171] 37 Control Department

[0172] 41 CPU

[0173] 43 Memory

[0174] 45 Storage Unit

[0175] 47 Input Section

[0176] 49 Output Section

[0177] 51 Ministry of Communications

[0178] 53 Network

[0179] 55 server

[0180] 61 Inherent Strain DB Production / Storage Department

[0181] 63 segmented sections

[0182] 65. Definition of Inherent Strain

[0183] 67 Parallel Computing Unit

[0184] 69 Calculation Results Comprehensive Department

[0185] 71 Frame

[0186] 73 Interior Design Department

[0187] 100 Analysis Device

[0188] DB Inherent Strain Database

[0189] Model for MDL Validation

[0190] W-shaped layered design.

Claims

1. A deformation prediction method of a layered molded product, which is a deformation prediction method of a layered molded product molded by repeatedly layering a bead of deposited weld formed of a filler material melted and solidified, wherein the deformation prediction method of the layered molded product comprises: a step of dividing a shape of the layered molded product into a plurality of blocks; a step of determining inherent strain, accumulating information of inherent strain in each case from analysis results under various passes and welding conditions and measured results of relative displacements of adjacent passes from each other, thereby constructing a database, and determining inherent strain corresponding to the blocks respectively with reference to the database; a step of calculating, by parallel processing of a plurality of threads, the relative deformation amount and the deformation direction of each of the blocks before and after the bead of deposited weld using the inherent strain corresponding to the blocks respectively based on the inherent strain method; a step of setting at least one block group composed of blocks that engage with each other among the plurality of blocks; and a step of calculating deformation of the block group as a whole by adding the deformation amounts of the blocks constituting the block group according to the deformation directions of the blocks.

2. The deformation prediction method of the layered molded product according to claim 1, wherein the step of dividing into the plurality of blocks divides the shape of the layered molded product in units of the bead of deposited weld.

3. The deformation prediction method of the layered molded product according to claim 1, wherein the step of dividing into the plurality of blocks divides the shape of the layered molded product in units of a layered body of a plurality of the bead of deposited weld.

4. The deformation prediction method of the layered molded product according to any one of claims 1 to 3, wherein blocks that become the same welding condition with each other among the plurality of blocks are set as the same block group as each other.

5. The deformation prediction method of the layered molded product according to any one of claims 1 to 3, wherein in a case where the shape of the layered molded product is divided into a plurality of the block groups, the deformation of the layered molded product is predicted by adding the deformation amounts of the block groups according to the deformation directions of the block groups.

6. The deformation prediction method of the layered molded product according to claim 4, wherein in a case where the shape of the layered molded product is divided into a plurality of the block groups, the deformation of the layered molded product is predicted by adding the deformation amounts of the block groups according to the deformation directions of the block groups.

Images