Process Adjustment Methods for Additive Manufacturing of Heterogeneous Materials

Abstract

The present invention provides a process adjustment method for additive manufacturing of heterogeneous materials, which is applied to the fields of additive manufacturing technology and computer technology. The method comprises: determining the first process parameter of the i-th heterogeneous material component transition layer according to the gradient design requirement parameter of the i-th heterogeneous material component transition layer; for the i-th heterogeneous material component transition layer, repeatedly performing the following steps until M printing points in the i-th heterogeneous material component transition layer are printed: obtaining the real-time cooling information of the m-th printing point of the heterogeneous material during the additive manufacturing process; inputting the first process parameter and the real-time cooling information of the m-th printing point into a crack prediction model to obtain the crack probability at the heterogeneous material interface of the m-th printing point; when the crack probability at the heterogeneous material interface of the m-th printing point meets the preset crack probability range, adjusting the second process parameter of the m-th printing point so as to print the m+1-th printing point using the second process parameter during the additive manufacturing process.

Description

Technical Field

[0001] The present invention relates to the technical field of additive manufacturing and the technical field of computer science, and more specifically, to a process adjustment method for additive manufacturing of heterogeneous materials. Background Art

[0002] Valves are important components in piping systems and are widely used in many industries such as ships, energy, and electricity, especially in the field of ships. There are a large number of valve structures on ships, and the application demand is very large. Many valves have been in service in harsh marine environments for a long time and are easily affected by the marine salt spray environment, causing corrosion and damage on the surface. Once a valve is damaged, it will cause a catastrophic accident, which places very high demands on the manufacturing quality of the valve.

[0003] In the related art, the main body of the copper-steel composite valve is usually made of high-strength steel, and a copper alloy of a certain thickness is manufactured on the outer surface of the valve. However, since the copper-steel composite valve is made of heterogeneous materials, the composition of the copper-steel dissimilar metal transition zone varies greatly during the valve manufacturing process, which is prone to problems such as poor bonding performance and easy damage. Summary of the invention

[0004] In view of this, the present invention proposes a process adjustment method for additive manufacturing of heterogeneous materials, so as to realize component control of the copper-steel dissimilar metal component transition zone in the process of manufacturing valves, achieve a relatively wide dissimilar metal component transition zone, and realize reliable connection of dissimilar metal areas.

[0005] One aspect of the present invention provides a process adjustment method for additive manufacturing of heterogeneous materials, comprising: determining a first process parameter of the i-th heterogeneous material component transition layer according to a gradient design requirement parameter of the i-th heterogeneous material component transition layer in a heterogeneous material component transition zone, wherein the i-th heterogeneous material component transition layer includes M printing points during the additive manufacturing process, and the first process parameter is a process parameter required when using an arc as a heat input to melt the heterogeneous material, wherein i is a positive integer ≥1;

[0006] For the i-th heterogeneous material component transition layer, the following steps are repeated until M printing points in the i-th heterogeneous material component transition layer are printed:

[0007] The real-time cooling information of the m-th printing point among the M printing points of the heterogeneous material in the additive manufacturing process is obtained, where M is a positive integer ≥1, and 1≤m≤M; the first process parameter and the real-time cooling information of the m-th printing point are input into a cracking prediction model to obtain the cracking probability at the heterogeneous material interface of the m-th printing point, wherein the cracking prediction model is used to predict the cracking probability of the transition layer of the heterogeneous material composition; when the cracking probability at the heterogeneous material interface of the m-th printing point satisfies a preset cracking probability range, the second process parameter of the m-th printing point is adjusted so as to print the m+1-th printing point using the second process parameter during the additive manufacturing process, wherein the second process parameter is the process parameter required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool using a laser.

[0008] Optionally, determining the first process parameters of the i-th heterogeneous material component transition layer according to the gradient design requirement parameters of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone includes: obtaining the gradient design requirement parameters of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone under preset conditions; calling a database storage component to determine the first process parameters of the i-th heterogeneous material component transition layer based on the correlation between the gradient design requirement parameters and the first process parameters.

[0009] Optionally, the preset conditions include at least one of the following: a preset load cycle variation curve, a preset temperature curve, and a preset service acid-base environment.

[0010] Optionally, obtaining real-time cooling information of the mth printing point among M printing points in the additive manufacturing process of the heterogeneous material includes: determining a real-time cooling curve of the mth printing point in the additive manufacturing process of the heterogeneous material; performing data processing on the real-time cooling curve to obtain real-time cooling information of the mth printing point among M printing points in the additive manufacturing process of the heterogeneous material.

[0011] Optionally, determining the real-time cooling curve of the mth printing point of the heterogeneous material in the additive manufacturing process includes: receiving the real-time temperature of the mth printing point of the heterogeneous material in the additive manufacturing process acquired by a temperature sensor; and determining the real-time cooling curve of the mth printing point of the heterogeneous material in the additive manufacturing process according to the real-time temperature of the mth printing point.

[0012] Optionally, the crack prediction model includes a trained first neural network and a trained second neural network.

[0013] Optionally, the first process parameter and the real-time cooling information of the m-th printing point are input into a crack prediction model to obtain the crack probability at the heterogeneous material interface of the m-th printing point, including: inputting the first process parameter and the real-time cooling information of the m-th printing point into a trained first neural network to obtain the grain size distribution matrix of the heterogeneous material of the m-th printing point; inputting the grain size distribution matrix of the heterogeneous material into a trained second neural network to obtain the crack probability at the heterogeneous material interface of the m-th printing point.

[0014] Optionally, the crack prediction model is trained by the following operations: obtaining a first training sample set, wherein the first training sample set includes first sample process parameters of a sample heterogeneous material component transition layer during an additive manufacturing process and sample real-time cooling information corresponding to each printing point; based on the first sample process parameters and sample real-time cooling information corresponding to each printing point, adjusting the parameters of the crack prediction model to be trained to obtain a trained crack prediction model.

[0015] Optionally, the crack prediction model to be trained includes a pre-trained first neural network and a pre-trained second neural network.

[0016] Optionally, based on the first sample process parameters and the sample real-time cooling information corresponding to each printing point, the parameters of the crack prediction model to be trained are adjusted, including: using a pre-trained first neural network to perform feature processing on the first sample process parameters and the sample real-time cooling information corresponding to each printing point to obtain a sample grain size distribution matrix corresponding to each printing point; using a pre-trained second neural network to perform transformation processing on the sample grain size distribution matrix corresponding to each printing point to obtain a predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; adjusting the parameters of the crack prediction model to be trained according to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point.

[0017] Optionally, the pre-trained first neural network is trained by the following operations: obtaining a second training sample set, wherein the second training sample set includes first sample process parameters of the sample heterogeneous material component transition layer during the additive manufacturing process and sample real-time cooling information corresponding to each printing point; performing feature processing on the first sample process parameters of the sample heterogeneous material component transition layer during the additive manufacturing process and the sample real-time cooling information corresponding to each printing point to obtain a sample grain size distribution prediction matrix corresponding to each printing point; adjusting the parameters of the first neural network to be pre-trained according to the sample grain size distribution prediction matrix and the grain size distribution true matrix corresponding to each printing point to obtain a trained first neural network.

[0018] Optionally, the pre-trained second neural network is trained by the following operations: obtaining a third training sample set, wherein the third training sample set includes a sample grain size distribution matrix corresponding to each printing point of the sample heterogeneous material component transition layer during the additive manufacturing process; transforming the sample grain size distribution matrix to obtain a predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; adjusting the parameters of the second neural network to be pre-trained according to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point to obtain a pre-trained second neural network.

[0019] Another aspect of the present invention provides an electronic device, comprising:

[0020] one or more processors;

[0021] a memory for storing one or more programs,

[0022] When one or more programs are executed by one or more processors, the one or more processors implement the above method.

[0023] Another aspect of the present invention provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the above method when executed.

[0024] Another aspect of the present invention provides a computer program product, the computer program product comprising computer executable instructions, and the instructions are used to implement the above method when being executed.

[0025] According to an embodiment of the present invention, by obtaining the first process parameters of each heterogeneous material composition transition layer required for arc melting of heterogeneous materials during the additive manufacturing process of heterogeneous materials, and obtaining real-time cooling information of each printing point in each heterogeneous material composition transition layer in real time, the cracking probability at the heterogeneous material interface of each printing point is determined, and based on the cracking probability, the process parameters required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool by laser in each printing point are adjusted in real time, so as to solve the problem that the composition of the copper-steel dissimilar metal composition transition zone varies greatly during the valve manufacturing process, and poor bonding performance and easy damage may occur. Furthermore, the process parameters required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool by laser can be adjusted in real time, so as to achieve the technical effect of realizing reliable connection of the metal areas of the heterogeneous materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0027] Figure 1A flow chart of a process adjustment method for heterogeneous material additive manufacturing according to an embodiment of the present invention is shown;

[0028] Figure 2 The metallographic microstructure of the copper-steel heterogeneous material interface is shown;

[0029] Figure 3 A schematic diagram showing the effect of additive manufacturing of copper-steel heterogeneous materials using a dual-energy beam eutectic pool according to an embodiment of the present invention is shown;

[0030] Figure 4 A schematic diagram showing the training process of the first neural network;

[0031] Figure 5 A schematic diagram of the network structure of the cracking prediction model is shown;

[0032] Figure 6 A block diagram of an electronic device suitable for implementing a process adjustment method for additive manufacturing of heterogeneous materials according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0033] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0034] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0035] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0036] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0037] Figure 1 A flow chart of a process adjustment method for additive manufacturing of heterogeneous materials according to an embodiment of the present invention is shown.

[0038] like Figure 1 As shown, the method includes operations S110 to S120.

[0039] In operation S110 , a first process parameter of the i th heterogeneous material composition transition layer is determined according to a gradient design requirement parameter of the i th heterogeneous material composition transition layer in the heterogeneous material composition transition zone, wherein i is a positive integer ≥1.

[0040] According to an embodiment of the present invention, in the additive manufacturing process of a heterogeneous material structural part, a heterogeneous material composition transition zone is generated by each generated heterogeneous material composition transition layer.

[0041] According to an embodiment of the present invention, the heterogeneous material composition transition layer may be formed by two or more materials with different compositions. The gradient design requirement parameters of each heterogeneous material composition transition layer may include the composition design of the heterogeneous material of each transition layer and the thickness design of the transition layer.

[0042] According to the embodiment of the present invention, due to the different thicknesses of the transition layer of heterogeneous material composition, the corresponding transition layer performance is different. The appropriate thickness is set according to actual needs. The smoother the transition of the heterogeneous material composition transition layer of the manufactured heterogeneous material structural part is, the better the corresponding mechanical performance is.

[0043] For example, taking copper-steel heterogeneous material structural parts as an example, Figure 2 Figure 2 shows the metallographic microstructure of the copper-steel heterogeneous material interface. Figure 2 As shown in the figure, since the diffusion coefficient of Fe in the Cu matrix is ​​much higher than that of Cu in the Fe matrix, the Fe in the molten pool is more inclined to diffuse to the aluminum bronze side to form a supersaturated solid solution of Fe. In the subsequent solidification process, the solubility of Fe in the aluminum bronze matrix decreases, thereby precipitating to form an iron-rich phase. As the molten pool solidifies, the temperature gradient of the molten pool will also change. The iron-rich phase will gradually grow from the original spherical and flower-shaped dendrite arms, and will present a dendrite shape after complete cooling. Due to the limited diffusion distance of Fe, the iron-rich phase is relatively rare in areas far from the interface. As it gets closer to the interface, the iron content gradually increases, the size of the iron-rich phase gradually increases, and the number of distributions gradually increases. In some areas, due to poor diffusion and distribution of Fe, there are penetration cracks at the interface. Therefore, copper-steel heterogeneous structural parts need to control the heterogeneous material composition and transition layer thickness of the heterogeneous material transition layer.

[0044] According to an embodiment of the present invention, taking the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone as an example, the first process parameters of the i-th heterogeneous material component transition layer can be obtained based on the gradient design requirement parameters of the i-th heterogeneous material component transition layer.

[0045] According to an embodiment of the present invention, the first process parameter may be a process parameter required for providing a larger heat input to the dual energy beam eutectic pool through the arc to melt the heterogeneous material. The first process parameter may include at least one of the arc swing amplitude, the arc swing frequency, and the arc swing trajectory.

[0046] According to an embodiment of the present invention, the i-th heterogeneous material composition transition layer may include M printing points during the additive manufacturing process, where M is a positive integer ≥1.

[0047] In operation S120, for the i-th heterogeneous material composition transition layer, the following steps are repeatedly performed until the M printing dots in the i-th heterogeneous material composition transition layer are printed. Operation S120 also includes: operations S120-1 to S120-3.

[0048] In operation S120 - 1 , real-time cooling information of an m-th printing point among M printing points of a heterogeneous material in an additive manufacturing process is obtained, where 1≤m≤M.

[0049] In operation S120 - 2 , the first process parameter and the real-time cooling information of the m th printing point are input into a crack prediction model to obtain a crack probability at a heterogeneous material interface of the m th printing point.

[0050] In operation S120-3, when the crack probability at the heterogeneous material interface of the mth printing point satisfies a preset crack probability range, the second process parameter of the mth printing point is adjusted so that the m+1th printing point is printed using the second process parameter during the additive manufacturing process.

[0051] According to an embodiment of the present invention, the real-time cooling information may represent the printing temperature of the mth printing point, and obtain the cooling rate of the mth printing point based on the printing temperature cooling process of the printing point.

[0052] According to an embodiment of the present invention, the crack prediction model is used to predict the crack probability of the transition layer of heterogeneous material components; the second process parameter is the process parameter required when using laser to control the element distribution between the heterogeneous materials after melting in the eutectic pool.

[0053] According to an embodiment of the present invention, the second process parameter may include at least one of laser swing amplitude, laser swing frequency, laser swing trajectory, laser arc spacing, and laser spot size.

[0054] According to an embodiment of the present invention, the mth printing point is taken as the first printing point of the i-th heterogeneous material composition transition layer, based on the first process parameter, the heterogeneous material is melted by an electric arc, and based on the initial second process parameter, the molten pool is stirred by a laser to achieve mixing and composition control of the heterogeneous materials, the printing temperature of the first printing point is obtained in real time, and the cooling rate of the first printing point is obtained based on the printing temperature cooling process of the first printing point, that is, the real-time cooling information of the first printing point.

[0055] According to an embodiment of the present invention, the first process parameter corresponding to the first printing point and the real-time cooling information of the first printing point acquired in real time are input into the crack prediction model, so as to obtain the crack probability at the heterogeneous material interface of the first printing point.

[0056] According to an embodiment of the present invention, when it is determined that the crack probability at the heterogeneous material interface of the first printing point meets the preset crack probability range, it means that the heterogeneous material interface of the first printing point is prone to cracking. Therefore, the initial second process parameters of the first printing point are adjusted to obtain the adjusted second process parameters of the first printing point, so as to print the second printing point using the adjusted second process parameters of the first printing point.

[0057] According to an embodiment of the present invention, an arc is used to melt heterogeneous materials based on a first process parameter, and based on an adjusted second process parameter of a first printing point, a laser is used to stir the molten pool, thereby achieving mixing and composition control of the heterogeneous materials, obtaining a printing temperature of a second printing point in real time, and obtaining a cooling rate of the second printing point based on a cooling process of the printing temperature of the second printing point, i.e., real-time cooling information of the second printing point.

[0058] According to an embodiment of the present invention, the first process parameter corresponding to the second printing point and the real-time cooling information of the second printing point acquired in real time are input into the crack prediction model, so as to obtain the crack probability at the heterogeneous material interface of the second printing point.

[0059] According to an embodiment of the present invention, when it is determined that the crack probability at the heterogeneous material interface of the second printing point satisfies a preset crack probability range, it indicates that the heterogeneous material interface of the second printing point is prone to cracking. Therefore, the second process parameters of the second printing point are adjusted to obtain the adjusted second process parameters of the second printing point, so as to print the third printing point using the adjusted second process parameters of the second printing point.

[0060] According to an embodiment of the present invention, when it is determined that the crack probability at the heterogeneous material interface of the second printing point does not meet the preset crack probability range, it means that the heterogeneous material interface of the second printing point is not prone to cracking, and the third printing point can continue to be printed using the adjusted second process parameters of the first printing point.

[0061] According to an embodiment of the present invention, based on the above method, an arc can be used to melt heterogeneous materials based on the first process parameters, and based on the adjusted second process parameters of the M-1th printing point, a laser can be used to stir the molten pool, thereby achieving mixing and composition control of the heterogeneous materials, and printing of the Mth printing point of the i-th heterogeneous material composition transition layer, thereby completing the generation of the i-th heterogeneous material composition transition layer.

[0062] According to an embodiment of the present invention, the above method can be used to realize the generation of the remaining heterogeneous material composition transition layers in the heterogeneous material composition transition zone, thereby generating the required thickness of the heterogeneous material composition transition zone.

[0063] For example, Figure 3 The schematic diagram of the effect of additive manufacturing of copper-steel heterogeneous materials using a dual-energy beam eutectic pool according to an embodiment of the present invention is shown. Figure 3 As shown, based on the first process parameter, an electric arc is used to provide a larger heat input to the dual-energy beam eutectic pool to melt the deposited heterogeneous material metal; and based on the second process parameter, a laser is used to stir the eutectic pool to achieve mixing and composition control of the heterogeneous material metal, to achieve a relatively wide heterogeneous material composition transition zone, and to achieve reliable connection of the copper-steel dissimilar metal area.

[0064] According to an embodiment of the present invention, by obtaining the first process parameters of each heterogeneous material composition transition layer required for arc melting of heterogeneous materials during the additive manufacturing process of heterogeneous materials, and obtaining real-time cooling information of each printing point in each heterogeneous material composition transition layer in real time, the cracking probability at the heterogeneous material interface of each printing point is determined, and based on the cracking probability, the process parameters required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool by laser in each printing point are adjusted in real time, so as to solve the problem that the composition of the copper-steel dissimilar metal composition transition zone varies greatly during the valve manufacturing process, and poor bonding performance and easy damage may occur. Furthermore, the process parameters required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool by laser can be adjusted in real time, so as to achieve the technical effect of realizing reliable connection of the metal areas of the heterogeneous materials.

[0065] According to an embodiment of the present invention, determining the first process parameters of the i-th heterogeneous material component transition layer according to the gradient design requirement parameters of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone includes: obtaining the gradient design requirement parameters of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone under preset conditions; calling a database storage component to determine the first process parameters of the i-th heterogeneous material component transition layer based on the correlation between the gradient design requirement parameters and the first process parameters.

[0066] According to an embodiment of the present invention, the preset conditions may include a preset load cycle variation curve, a preset temperature curve or a preset service acid-base environment.

[0067] According to an embodiment of the present invention, the database storage component can call a database for storing first process parameters. Based on the obtained gradient design requirement parameters of the heterogeneous material composition transition zone under preset conditions, the first process parameters corresponding to the gradient design requirement parameters of each heterogeneous material composition transition zone can be searched from the database.

[0068] According to an embodiment of the present invention, the first process parameters corresponding to the gradient design requirement parameters of each heterogeneous material composition transition zone can be characterized and stored in a database in a coded manner.

[0069] According to an embodiment of the present invention, real-time cooling information of the mth printing point among M printing points in an additive manufacturing process of a heterogeneous material is obtained, including: determining a real-time cooling curve of the mth printing point in the additive manufacturing process of the heterogeneous material; performing data processing on the real-time cooling curve to obtain real-time cooling information of the mth printing point among M printing points in the additive manufacturing process of the heterogeneous material.

[0070] According to an embodiment of the present invention, determining a real-time cooling curve of an m-th printing point of a heterogeneous material in an additive manufacturing process includes: receiving a real-time temperature of the m-th printing point of the heterogeneous material in the additive manufacturing process acquired by a temperature sensor; and determining a real-time cooling curve of the m-th printing point of the heterogeneous material in the additive manufacturing process based on the real-time temperature of the m-th printing point.

[0071] According to an embodiment of the present invention, the printing temperature of the mth printing point of a heterogeneous material structure in an additive manufacturing process can be collected in real time based on a temperature sensor, and the printing temperature of the mth printing point can be sent to a computer device.

[0072] According to an embodiment of the present invention, the temperature sensor may be a contact type or a non-contact type temperature sensor.

[0073] According to an embodiment of the present invention, a real-time cooling curve of the heterogeneous material deposited metal may be plotted point by point based on the cooling process of the received real-time printing temperature of the mth printing point.

[0074] According to an embodiment of the present invention, the real-time cooling curve represents a curve of the real-time printing temperature of the heterogeneous material deposited metal cooling over time.

[0075] According to an embodiment of the present invention, the maximum printing temperature of the mth printing point can be determined from the real-time cooling curve, and the tangent corresponding to the equilibrium temperature point can be drawn based on the real-time cooling curve, and the real-time cooling rate of the mth printing point can be determined based on the slope of the tangent.

[0076] According to an embodiment of the present invention, the maximum printing temperature and the real-time cooling rate of the m-th printing point may be used as the real-time cooling information of the m-th printing point.

[0077] According to an embodiment of the present invention, the first process parameter and the real-time cooling information of the m-th printing point are input into a crack prediction model to obtain the crack probability at the heterogeneous material interface of the m-th printing point, including: inputting the first process parameter and the real-time cooling information of the m-th printing point into a trained first neural network to obtain the grain size distribution matrix of the heterogeneous material of the m-th printing point; inputting the grain size distribution matrix of the heterogeneous material into a trained second neural network to obtain the crack probability at the heterogeneous material interface of the m-th printing point.

[0078] According to an embodiment of the present invention, the crack prediction model may include a trained first neural network and a trained second neural network.

[0079] According to an embodiment of the present invention, the first process parameters of the i-th heterogeneous material composition transition layer and the real-time cooling information of the m-th printing point are input into the trained first neural network of the crack prediction model, and the grain size distribution matrix of the heterogeneous material of the m-th printing point can be obtained.

[0080] According to an embodiment of the present invention, the grain size distribution matrix can characterize the mechanical properties of different regions in the transition layer of heterogeneous material composition.

[0081] According to an embodiment of the present invention, the grain size distribution matrix of the mth printing point can be input into the trained second neural network of the crack prediction model, and the grain size distribution matrix of the mth printing point can be converted into the crack probability of the corresponding heterogeneous material, thereby predicting the crack probability at the heterogeneous material interface of the mth printing point.

[0082] According to an embodiment of the present invention, a crack prediction model can be used to perform feature processing on the first process parameter and the real-time cooling information of the current printing point acquired in real time, predict the crack probability at the heterogeneous material interface of the current printing point, and adjust the second process parameter for controlling the heterogeneous material composition using a laser based on the crack probability so as to print the next printing point, thereby realizing the redistribution of the heterogeneous material composition, achieving reliable connection of the heterogeneous material transition zone, and improving the mechanical properties of heterogeneous material additively manufactured structural parts.

[0083] According to an embodiment of the present invention, the crack prediction model is obtained by adjusting parameters of the crack prediction model to be trained.

[0084] According to an embodiment of the present invention, the crack prediction model to be trained includes a pre-trained first neural network and a pre-trained second neural network. The first neural network may include several convolutional layers, and the second neural network may include several convolutional layers and one fully connected layer.

[0085] According to an embodiment of the present invention, the pre-trained first neural network is obtained by the following operation training method: obtaining a second training sample set; performing feature processing on the first sample process parameters of the sample heterogeneous material component transition layer in the additive manufacturing process and the sample real-time cooling information corresponding to each printing point to obtain the sample grain size distribution prediction matrix corresponding to each printing point; adjusting the parameters of the first neural network to be pre-trained according to the sample grain size distribution prediction matrix and the grain size distribution real matrix corresponding to each printing point to obtain the pre-trained first neural network.

[0086] According to an embodiment of the present invention, the second training sample set may include first sample process parameters of a sample heterogeneous material component transition layer during additive manufacturing and sample real-time cooling information corresponding to each printing point in the transition layer.

[0087] According to an embodiment of the present invention, the first sample process parameters of the sample heterogeneous material composition transition layer in the additive manufacturing process and the sample real-time cooling information corresponding to each printing point in the transition layer can be input into the first neural network to be pre-trained for forward propagation, and the sample grain size distribution prediction matrix H1 corresponding to each printing point can be obtained, and the grain size distribution real matrix corresponding to each printing point can be obtained based on the real matrix Constructing the loss function A loss value is obtained, and based on the loss value, back propagation is performed to adjust the parameters of the first neural network to be pre-trained, until the preset conditions of the iteration are met, the training can be terminated to obtain the pre-trained first neural network. The preset conditions of the iteration can be that the loss value reaches convergence or the performance index of the first neural network meets the threshold requirement.

[0088] According to an embodiment of the present invention, through forward propagation, the feature data of the second training sample set can be processed layer by layer through each convolution layer of the first neural network to be pre-trained to obtain a prediction result; back propagation compares the prediction result of the second training sample set with the actual result to determine the update amplitude of each convolution kernel weight of the first neural network to be pre-trained, that is, back propagation is used to determine the change of the loss function relative to each convolution kernel weight, which can also be called a gradient or error derivative, recorded as ,in, Represents the weight of mapping convolution kernel l to convolution kernel n.

[0089] For example, Figure 4 A schematic diagram of the training process of the first neural network is shown. Figure 4 As shown, the first neural network includes three convolution layers and six convolution kernels as an example. The operation at each convolution kernel is similar to that of convolution kernel 6, and the following two formulas can be used to describe the forward propagation process of the first neural network:

[0090] (1);

[0091] (2);

[0092] Where y represents the input data of the activation function f(·) of the convolution kernel; z represents the output of the convolution kernel; the subscript n or l represents the serial number of the convolution kernel; in(n) represents the set of serial numbers of the convolution kernels in the previous layer of convolution kernel n, for example, Figure 4 In the figure, convolution kernel 4 receives the outputs of convolution kernel 1, convolution kernel 2, and convolution kernel 3, in(4)={1,2,3}; Represents the weight of mapping convolution kernel l to convolution kernel n; Represents the constant term corresponding to the convolution kernel n. Among them, and The parameters that make up the first neural network can be obtained through training.

[0093] like Figure 4 As shown, we can first calculate the gradient of the loss function value relative to the output of the output convolution kernel 6 , when the loss function is the mean square error loss function hour, ,in, To predict the results, is the real result. Then, the chain rule can be used to calculate the weight of the loss function value relative to the mapping of convolution kernel 5 to convolution kernel 6. Gradient And the gradient of the loss function relative to the output of the output convolution kernel 5 ,Right now:

[0094] (3);

[0095] (4);

[0096] (5);

[0097] By analogy, the gradient of the loss function value relative to each weight can be calculated one by one. Based on the above process, the gradient can be back-propagated based on the loss function value until the gradient of the loss function value relative to each element in the initial output matrix is ​​calculated, and then the parameters are updated based on the gradient of each element, thereby realizing a round of updates of the first neural network. Multiple rounds of iterations are implemented based on the above algorithm until the loss value converges or the performance index of the first neural network meets the threshold requirement, thereby obtaining a pre-trained first neural network.

[0098] According to an embodiment of the present invention, the pre-trained second neural network is trained by the following operations: obtaining a third training sample set; transforming the sample grain size distribution matrix to obtain the predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; adjusting the parameters of the second neural network to be pre-trained according to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point to obtain the pre-trained second neural network.

[0099] According to an embodiment of the present invention, the third training sample set may include a sample grain size distribution matrix corresponding to each printing point of a sample heterogeneous material composition transition layer during an additive manufacturing process.

[0100] According to an embodiment of the present invention, the sample grain size distribution matrix corresponding to each printing point can be input into the second neural network to be pre-trained for forward propagation, and the predicted cracking probability corresponding to each printing point can be obtained. A loss value is obtained based on the actual cracking probability and the predicted cracking probability using a loss function, and back propagation is performed based on the loss value to adjust the parameters of the second neural network to be pre-trained until the iterative preset condition is met to terminate the training and obtain the pre-trained second neural network. The iterative preset condition can be that the loss value reaches convergence or the performance index of the second neural network meets the threshold requirement.

[0101] It should be noted that the forward propagation and back propagation processes of the second neural network during the training process are the same as the forward propagation and back propagation processes of the first neural network during the training process, and the present invention will not be repeated here.

[0102] According to an embodiment of the present invention, the pre-trained first neural network and the pre-trained second neural network can be used separately according to actual conditions to meet the needs of different usage scenarios. For example, the pre-trained first neural network can couple the relationship between additive manufacturing parameters and grain size distribution, and then predict the tensile strength, yield strength, ultimate tensile strength, fracture toughness and other mechanical properties of the material based on the grain size distribution.

[0103] In some embodiments, the first neural network and the second neural network can also be models trained in other application scenarios, which are called in the current use scenario and trained in different ways. For example, the first neural network corresponding to a certain aluminum alloy has been obtained in other application scenarios, and the data of the current training scenario is used to fine-tune the parameters to adapt to the hyperparameter changes under the current printing material (copper-steel heterogeneous materials). Through such a design, the demand for training samples can be greatly reduced, and the number of training iterations and training time can be reduced.

[0104] In some embodiments, the model architecture of the second neural network can be set based on the knowledge of metallurgy. Based on metal damage mechanics, the probability of cracking of metal under a certain load can be obtained based on the grain size distribution. Exemplarily, the coupling relationship between grain size distribution and metal cracking behavior can be established by finite element simulation analysis software through intracrystalline plastic finite element method, and then the architecture and hyperparameters of the second neural network can be designed based on this relationship. Through such a design, the demand for samples can be effectively reduced, and the difficulty of convergence of training samples can be greatly reduced.

[0105] According to an embodiment of the present invention, the crack prediction model is obtained by jointly training a pre-trained first neural network and a pre-trained second neural network. The training process includes: obtaining a first training sample set; adjusting the parameters of the crack prediction model to be trained based on the first sample process parameters and the real-time cooling information of the samples corresponding to each printing point, and obtaining a trained crack prediction model.

[0106] According to an embodiment of the present invention, the first training sample set may include first sample process parameters of a sample heterogeneous material component transition layer during an additive manufacturing process and sample real-time cooling information corresponding to each printing point.

[0107] According to an embodiment of the present invention, based on the first sample process parameters and the sample real-time cooling information corresponding to each printing point, the parameters of the crack prediction model to be trained are adjusted, including: using a pre-trained first neural network to perform feature processing on the first sample process parameters and the sample real-time cooling information corresponding to each printing point to obtain a sample grain size distribution matrix corresponding to each printing point; using a pre-trained second neural network to perform transformation processing on the sample grain size distribution matrix corresponding to each printing point to obtain a predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; adjusting the parameters of the crack prediction model to be trained according to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point.

[0108] According to an embodiment of the present invention, the first sample process parameters of the sample heterogeneous material component transition layer in the additive manufacturing process and the sample real-time cooling information corresponding to each printing point can be input into the pre-trained first neural network of the crack prediction model for forward propagation, and the sample grain size distribution matrix corresponding to each printing point can be obtained. The sample grain size distribution matrix is ​​then input into the pre-trained second neural network of the crack prediction model for forward propagation to obtain the predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; based on the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point, a loss value is obtained, and back propagation is performed based on the loss value to adjust the parameters of the crack prediction model to be trained until the preset iteration condition is reached, and the training is terminated to obtain a trained crack prediction model.

[0109] According to an embodiment of the present invention, the crack prediction model can adopt a multi-step convolutional neural network, and the training process of different convolutional layers can be multi-threaded and trained in parallel. The training process can also be performed on different host terminals, which greatly reduces the time required for the training process and greatly reduces the hardware cost required for the training process.

[0110] Figure 5 Figure 2 shows a schematic diagram of the network structure of the crack prediction model. Figure 5 As shown, the first process parameters corresponding to the heterogeneous material component transition layer in the first training sample set and the sample real-time cooling information 1, ..., the first process parameters and the sample real-time cooling information n corresponding to each printing point can be input into the first neural network of the crack prediction model for feature processing. The first neural network can include convolution layers 1, ..., convolution layers i, and obtain the sample grain size distribution matrix corresponding to each printing point, that is, grain size distribution matrix 1, ..., grain size distribution matrix n; then the sample grain size distribution matrix corresponding to each printing point is input into the second neural network of the crack prediction model. The second neural network can include convolution layers 1, ..., convolution layers j, and a fully connected layer, and outputs the predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point, that is, cracking probability 1, ..., cracking probability n.

[0111] Figure 6 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present invention is shown. Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0112] like Figure 6 As shown, the electronic device according to an embodiment of the present invention includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 to a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include an onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0113] In RAM 603, various programs and data required for the operation of electronic device 600 are stored. Processor 601, ROM 602 and RAM 603 are connected to each other via bus 604. Processor 601 performs various operations of the method flow according to the embodiment of the present invention by executing the programs in ROM 602 and / or RAM 603. It should be noted that the program can also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 can also perform various operations of the method flow according to the embodiment of the present invention by executing the programs stored in one or more memories.

[0114] According to an embodiment of the present invention, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the input / output (I / O) interface 605: an input portion 606 including a keyboard, a mouse, etc.; an output portion 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 608 including a hard disk, etc.; and a communication portion 609 including a network interface card such as a LAN card, a modem, etc. The communication portion 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage portion 608 as needed.

[0115] According to an embodiment of the present invention, the method flow according to an embodiment of the present invention can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, the above-mentioned functions defined in the system of the embodiment of the present invention are executed. According to an embodiment of the present invention, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0116] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0117] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device.

[0118] For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 602 and / or the RAM 603 described above and / or one or more memories other than the ROM 602 and the RAM 603 .

[0119] An embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the method provided by the embodiment of the present invention.

[0120] When the computer program is executed by the processor 601, the above functions defined in the system / device of the embodiment of the present invention are executed. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0121] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 609, and / or installed from a removable medium 611. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0122] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways, even if such a combination or combination is not explicitly recorded in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways. All these combinations and / or combinations fall within the scope of the present invention.

[0124] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination. The present invention does not depart from the scope of the present invention, and those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A process adjustment method for additive manufacturing of heterogeneous materials, characterized in that: The method comprises: Determine, according to the gradient design requirement parameter of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone, the first process parameter of the i-th heterogeneous material component transition layer, wherein the i-th heterogeneous material component transition layer includes M printing points in the additive manufacturing process, and the first process parameter is the process parameter required when the electric arc is used as the heat input to melt the heterogeneous material, wherein i is a positive integer ≥1; For the i-th heterogeneous material component transition layer, the following steps are repeatedly performed until the M printing points in the i-th heterogeneous material component transition layer are printed: Acquire real-time cooling information of an m-th printing point among the M printing points of the heterogeneous material in the additive manufacturing process, where M is a positive integer ≥1, and 1≤m≤M; Inputting the first process parameter and the real-time cooling information of the m-th printing point into a crack prediction model to obtain the crack probability at the heterogeneous material interface of the m-th printing point, wherein the crack prediction model is used to predict the crack probability of the heterogeneous material component transition layer, and the crack prediction model includes a trained first neural network and a trained second neural network; When the cracking probability at the interface of the heterogeneous materials of the mth printing point satisfies a preset cracking probability range, adjusting the second process parameter of the mth printing point so as to print the (m+1)th printing point using the second process parameter during additive manufacturing, wherein the second process parameter is a process parameter required for controlling the element distribution between the heterogeneous materials after melting in the eutectic pool using a laser; The step of inputting the first process parameter and the real-time cooling information of the m-th printing point into a crack prediction model to obtain the crack probability at the heterogeneous material interface of the m-th printing point includes: Inputting the first process parameter and the real-time cooling information of the m-th printing point into the trained first neural network to obtain a grain size distribution matrix of the heterogeneous material of the m-th printing point; The grain size distribution matrix of the heterogeneous material is input into the trained second neural network to obtain the cracking probability at the heterogeneous material interface of the mth printing point.

2. The method according to claim 1, characterized in that The step of determining the first process parameter of the i-th heterogeneous material component transition layer according to the gradient design requirement parameter of the i-th heterogeneous material component transition layer in the heterogeneous material component transition zone comprises: Obtaining the gradient design requirement parameters of the i-th heterogeneous material composition transition layer in the heterogeneous material composition transition zone under preset conditions; The database storage component is called to determine the first process parameter of the i-th heterogeneous material composition transition layer based on the correlation between the gradient design requirement parameter and the first process parameter.

3. The method according to claim 2, characterized in that The preset conditions include at least one of the following: a preset load cycle variation curve, a preset temperature curve, and a preset service acid-base environment.

4. The method according to claim 1, characterized in that The obtaining of real-time cooling information of the mth printing point among the M printing points of the heterogeneous material in the additive manufacturing process includes: Determining a real-time cooling curve of the mth printing point of the heterogeneous material during the additive manufacturing process; Data processing is performed on the real-time cooling curve to obtain real-time cooling information of the mth printing point among the M printing points in the additive manufacturing process of the heterogeneous material.

5. The method according to claim 4, characterized in that Determining the real-time cooling curve of the mth printing point of the heterogeneous material during the additive manufacturing process includes: Receiving a real-time temperature of the mth printing point of the heterogeneous material during the additive manufacturing process acquired by a temperature sensor; According to the real-time temperature of the m-th printing point, a real-time cooling curve of the m-th printing point of the heterogeneous material during the additive manufacturing process is determined.

6. The method according to claim 1, characterized in that The cracking prediction model is trained by the following operations: Acquire a first training sample set, wherein the first training sample set includes first sample process parameters of a sample heterogeneous material component transition layer in an additive manufacturing process and sample real-time cooling information corresponding to each printing point; Based on the first sample process parameters and the sample real-time cooling information corresponding to each printing point, the parameters of the crack prediction model to be trained are adjusted to obtain a trained crack prediction model.

7. The method according to claim 6, characterized in that The crack prediction model to be trained includes a pre-trained first neural network and a pre-trained second neural network; The adjusting of the parameters of the crack prediction model to be trained based on the first sample process parameters and the sample real-time cooling information corresponding to each printing point includes: Using the pre-trained first neural network, feature processing is performed on the first sample process parameter and the sample real-time cooling information corresponding to each printing point to obtain a sample grain size distribution matrix corresponding to each printing point; The pre-trained second neural network is used to transform the sample grain size distribution matrix corresponding to each printing point to obtain the predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; According to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point, the parameters of the cracking prediction model to be trained are adjusted.

8. The method according to claim 7, characterized in that The pre-trained first neural network is trained by the following operations: Acquire a second training sample set, wherein the second training sample set includes first sample process parameters of the sample heterogeneous material component transition layer during the additive manufacturing process and sample real-time cooling information corresponding to each printing point; Performing feature processing on the first sample process parameters of the sample heterogeneous material composition transition layer during the additive manufacturing process and the sample real-time cooling information corresponding to each printing point to obtain a sample grain size distribution prediction matrix corresponding to each printing point; According to the sample grain size distribution prediction matrix and the grain size distribution real matrix corresponding to each printing point, the parameters of the first neural network to be pre-trained are adjusted to obtain the pre-trained first neural network.

9. The method according to claim 7, characterized in that: The pre-trained second neural network is trained by the following operations: Acquire a third training sample set, wherein the third training sample set includes the sample grain size distribution matrix corresponding to each printing point of the sample heterogeneous material component transition layer during the additive manufacturing process; Transforming the sample grain size distribution matrix to obtain the predicted cracking probability at the sample heterogeneous material interface corresponding to each printing point; According to the predicted cracking probability and the actual cracking probability at the sample heterogeneous material interface corresponding to each printing point, the parameters of the second neural network to be pre-trained are adjusted to obtain the pre-trained second neural network.

Images