Simulation method, device and system for mechanical properties of dot matrix structure for additive manufacturing

Abstract

This application relates to a method, apparatus, and system for simulating the mechanical properties of lattice structures for additive manufacturing. The method includes: measuring pixels from a 3D object image to generate the forming rod diameter; using machine learning to train and fit the tilt angle-rod diameter, predicting the forming rod diameter based on design values, and establishing a lattice structure model using the predicted values; conducting compression tests and using machine learning to train and fit the diameter-mechanical parameters to establish a mechanical test model, predicting the elastic modulus of the average forming diameter; establishing a finite element simulation model and inputting the predicted elastic modulus value into the finite element simulation model for mechanical property analysis; and adjusting and fitting the finite element simulation model based on the generated mechanical property analysis results. This method can experimentally measure the forming dimensional deviation of the 3D object and the forming quality of the current process, and then use finite element software to compensate for the macroscopic mechanical deviations caused by the size effect of the microstructure, thus achieving mechanical property simulation.

Description

Technical Field

[0001] This application relates to the field of finite element numerical simulation technology, and more specifically to a method, apparatus, system and electronic equipment for simulating the mechanical properties of lattice structures for additive manufacturing. Background Technology

[0002] Because the laser spot diameter and scanning layer thickness in laser additive manufacturing do not approach zero, as the laser spot moves on the powder bed, the surface of the part exists in the form of overlapping melt channels, resulting in peaks and valleys with convex centers and concave sides between the melt channels, forming a step effect.

[0003] Because the diameter of the rods in a lattice structure is usually small (on the millimeter scale) and the units are densely arranged, the fabrication process is greatly affected by the forming accuracy of the printing equipment. This effect causes the melt channel to have a significant impact on dimensional deviations, and the final formed sample exhibits a clear step effect. It is precisely because of this step effect that there is a large difference between the simulation calculation results and the mechanical experimental results. Summary of the Invention

[0004] This application provides a method, apparatus, and system for simulating the mechanical properties of lattice structures for additive manufacturing. It can experimentally measure the forming size deviation of a three-dimensional object and the forming quality of the current process, and then compensate for the macroscopic mechanical deviation caused by the size effect of the microstructure through secondary development of finite element software, thereby realizing mechanical property simulation.

[0005] In a first aspect, this application provides a simulation method for the mechanical properties of lattice structures for additive manufacturing, comprising: measuring pixels based on an image of a formed three-dimensional object containing various diameters and support inclination angles to generate the formed rod diameter; using the rod diameter and inclination angle as inputs and the formed rod diameter as output, training the model using a machine learning artificial neural network method to fit the inclination angle-rod diameter, and then predicting the formed rod diameter based on the design value, and using the predicted value to establish a lattice structure model; conducting compression tests on three-dimensional compression mechanical specimens of different diameters to obtain mechanical parameters of different diameters, and training the model using a machine learning artificial neural network method to fit the diameter-mechanical parameters, establishing a mechanical test model, and then predicting the elastic modulus of the average formed diameter; establishing a finite element simulation model based on the lattice structure model and the mechanical test model, inputting the predicted elastic modulus value as a material parameter into the finite element simulation model, performing mechanical property analysis, and then adjusting and fitting the finite element simulation model based on the generated mechanical property analysis results.

[0006] In one alternative embodiment of the first aspect, the method for generating the forming rod diameter includes: forming a three-dimensional object comprising multiple diameters and support inclination angles using additive manufacturing technology; photographing the three-dimensional object using a light microscope to generate a three-dimensional object image comprising multiple diameters and support inclination angles; binarizing the three-dimensional object image and obtaining the pixel diameter of the forming rod; and converting the pixel diameter into the actual diameter according to the pixel resolution of the light microscope to generate the forming rod diameter.

[0007] In one alternative embodiment of the first aspect, when training using a machine learning artificial neural network method, the method includes: importing a set of modules for data processing, modeling, and evaluation; loading data containing bar diameter and inclination angle features and normalizing the data to generate a first dataset containing a training set and a test set; constructing a neural network regression model using a multilayer perceptron to fit the inclination angle-bar diameter and setting the neural network parameters and iteration number, and then training the neural network regression model using the training set; using the test set to predict the trained neural network regression model and calculating the mean square error and coefficient of determination, and evaluating the model; predicting the formed bar diameter based on the design value using the neural network regression model, and then using the predicted value to establish a lattice structure model.

[0008] In one alternative embodiment of the first aspect, when obtaining mechanical parameters for different diameters, the method includes: forming three-dimensional compression mechanical specimens of different diameters using additive manufacturing technology; conducting mechanical compression tests on the three-dimensional compression mechanical specimens of different diameters using a material testing device and recording the experimental results including compression load and displacement values; calculating the elastic modulus based on the experimental results and generating mechanical data based on the diameter of each three-dimensional compression mechanical specimen and the matching elastic modulus.

[0009] In one alternative embodiment of the first aspect, when training using a machine learning artificial neural network method, the method includes: importing a set of modules for data processing, modeling, and evaluation; loading data containing diameter and elastic modulus features and normalizing the data to generate a second dataset containing a training set and a test set; constructing a neural network regression model using a multilayer perceptron to fit diameter-mechanical parameters and setting the neural network parameters and iteration number, and then training the neural network regression model using the training set; using the test set to predict the trained neural network regression model and calculating the mean square error and coefficient of determination, and evaluating the model; setting the neural network regression model that meets the preset requirements as a mechanical test model, and then predicting the elastic modulus of the average forming diameter based on the mechanical test model.

[0010] In one alternative of the first aspect, when importing libraries for implementing data processing, modeling, and evaluation, the method includes: importing a set of modules that provide data reading and processing; importing a set of modules that provide machine learning algorithms and tools; and importing a set of modules for numerical computation.

[0011] In one alternative of the first aspect, when training the neural network regression model using a training set, the method includes: using the average error as the loss function of the neural network regression model; setting the hidden layers and the number of neurons they contain, and setting the maximum number of training iterations; and optimizing the model parameters using backpropagation and gradient descent methods.

[0012] In one alternative to the first aspect, the method for establishing the finite element simulation model includes:

[0013] Set material properties; set dynamic analysis parameters, including analysis step size, mass scaling, target time increment, field output, and history output type; mesh the lattice structure model and assemble it with a mechanical test model to form a finite element simulation model; define the external loads and boundary conditions experienced by the structure in the dynamic analysis.

[0014] In one alternative to the first aspect, when setting material properties, the method includes: setting basic material properties, including material density, elastic modulus, and Poisson's ratio; and setting plastic properties of the material, including stress and strain.

[0015] Set the rate dependence property of the material; set the material damage properties, including damage parameters and damage evolution.

[0016] In one alternative embodiment of the first aspect, when adjusting and fitting the finite element simulation model based on the generated mechanical property analysis results, the method includes: comparing the generated mechanical property analysis results with actual experimental data to generate error results; performing error analysis based on the error results and defining an error function based on the difference between the quantified experimental values ​​and the simulated values; adjusting the parameters of the finite element simulation model based on the error function; and iteratively optimizing the finite element simulation model to minimize the error function.

[0017] Secondly, this application provides a simulation device for the mechanical properties of lattice structures for additive manufacturing using the above-mentioned simulation method, comprising: an image processing module for measuring pixels based on an image of a formed three-dimensional object containing multiple diameters and support inclination angles, generating the formed rod diameter; a first training module for using the rod diameter and inclination angle as inputs, the formed rod diameter as output, and training using a machine learning artificial neural network method to fit the inclination angle-rod diameter, thereby predicting the formed rod diameter based on the design value, and using the predicted value to establish a lattice structure model; a second training module for performing compression tests on three-dimensional compression mechanical specimens of different diameters, obtaining mechanical parameters of different diameters, and using a machine learning artificial neural network method to train and fit the diameter-mechanical parameters, establishing a mechanical test model, and thereby predicting the elastic modulus of the average formed diameter; and a finite element simulation module for establishing a finite element simulation model based on the lattice structure model and the mechanical test model, inputting the predicted elastic modulus value as a material parameter into the finite element simulation model, performing mechanical property analysis, and then adjusting and fitting the finite element simulation model based on the generated mechanical property analysis results.

[0018] Thirdly, this application provides a simulation system for the mechanical properties of lattice structures for additive manufacturing, including the above-mentioned simulation device, and further includes: additive manufacturing equipment for forming three-dimensional objects and three-dimensional compression mechanical specimens using additive manufacturing technology; optical mirror equipment for capturing images; and material testing equipment for conducting material mechanical tests.

[0019] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0020] The accompanying drawings, which are incorporated herein and form part of this specification, illustrate one or more embodiments of the present application and, together with the description, serve to explain the principles of the present application and to enable those skilled in the art to make and use the present application.

[0021] Figure 1 It is based on existing physical objects of rods with different inclination angles and enlarged structural schematic diagrams.

[0022] Figure 2 This is a flowchart illustrating an exemplary method for simulating the mechanical properties of lattice structures for additive manufacturing, according to some embodiments of this application.

[0023] Figure 3 This is an exemplary schematic diagram of a specimen including various diameters and support inclination angles according to some embodiments of this application.

[0024] Figure 4This is an exemplary schematic diagram illustrating the conversion from a design model to a finite element simulation model based on a predicted value of the forming rod diameter, according to some embodiments of this application.

[0025] Figure 5 This is a schematic diagram of the compression process of an exemplary three-dimensional compressive mechanics specimen according to some embodiments of this application.

[0026] Figure 6 This is a schematic diagram of an exemplary mechanical simulation result according to some embodiments of this application.

[0027] Figure 7 This is a schematic diagram of an exemplary mechanical experiment result according to some embodiments of this application.

[0028] Figure 8 This is a schematic diagram illustrating an exemplary mechanical experiment and simulation comparison according to some embodiments of this application.

[0029] Figure 9 This is a schematic diagram of the kinetic energy-internal energy curve of an exemplary simulation process according to some embodiments of this application.

[0030] Figure 10 This is a schematic diagram comparing an exemplary simulation method of this application with experimental results according to some embodiments of this application.

[0031] Figure 11 This is a schematic diagram comparing an exemplary simulation method of this application with an unmodified simulation according to some embodiments of this application.

[0032] Figure 12 This is a comparative schematic diagram of an exemplary simulation method, experimental results, and uncorrected simulation based on some embodiments of this application.

[0033] Figure 13 This is a connection diagram of an exemplary simulation system for the mechanical properties of lattice structures for additive manufacturing, according to some embodiments of this application.

[0034] Figure 14 This is a schematic diagram of the structure of an exemplary electronic device according to some embodiments of this application. Detailed Implementation

[0035] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make this application more complete and comprehensive, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a full understanding of embodiments of this application.

[0036] Additive manufacturing is an integrated manufacturing technology that builds three-dimensional objects by stacking materials layer by layer. This technology uses heat sources such as lasers to melt powder or filaments, increasing the density of parts and improving material utilization.

[0037] A lattice structure is a topologically ordered three-dimensional structure composed of multiple repeating unit cells arranged periodically in space. This structure has excellent specific strength, specific stiffness, energy absorption and vibration reduction properties, and has been theoretically proven to be an effective structure that can significantly reduce the relative density of components while ensuring reliable mechanical performance.

[0038] Finite element method (FEM) numerical simulation refers to the approximate solution of differential equations by discretizing a continuous object or system into a finite number of elements. It is used to predict and analyze the behavior of complex physical systems in fields such as structural mechanics, heat conduction, fluid dynamics, and electromagnetics. The advantage of FEM lies in its ability to handle complex geometries, boundary conditions, and nonlinear material behavior. Currently, many mature FEM software programs are used in academic research and engineering applications, such as ABAQUS. In the simulation of lattice structures, FEM can be used to evaluate the mechanical properties of different lattice structures, such as elastic modulus and compressive strength. Through secondary development of ABAQUS, a large number of different structures can be efficiently evaluated, greatly improving efficiency and accuracy.

[0039] To ensure the accuracy of simulation calculations during the manufacturing process of lattice structures, a three-dimensional model that fully considers the processing precision of the equipment can be introduced and established. This model should accurately reflect the geometric deviations encountered by the lattice structure during the forming stage due to processing conditions, such as powder adhesion, so that an accurate simulation model can be established in the early design stage. In the laser cladding process of lattice structures, due to the operational limitations and process characteristics of the forming equipment, the rods of the lattice structure often exhibit geometric dimensional deviations due to different processing conditions, mainly in terms of the diameter of the rods. These deviations do not occur randomly; they are closely related to the relative position of the rods with respect to the laser beam.

[0040] When the inclination angle of the rod approaches 90 degrees, it means that the rod axis is perfectly parallel to the laser beam direction. Under these ideal conditions, the energy concentration of the laser beam is the highest, and the local heating of the rod is the most uniform, so the diameter of the rod can usually reach its maximum value. However, as the inclination angle of the rod gradually increases, the energy distribution of the laser beam on the cross-section of the rod will shift and diffuse, resulting in a decrease in heat transfer efficiency, which in turn compresses the actual solidification diameter of the rod.

[0041] When the rod is perpendicular to the laser beam, the energy utilization rate is lowest because the laser energy cannot directly act on the normal direction of the rod. Therefore, in this extreme case, the actual diameter of the rod will be reduced to a minimum. Simultaneously, when the rod diameter is small, the support rod is more susceptible to the influence of the SLM manufacturing process, making the error between the formed rod diameter and the designed rod diameter more significant. This phenomenon is due to the unequal melt depth and width of the laser melt pool during the printing process. When the rod inclination angle is small, the rod diameter is mainly affected by the melt width; while when the rod inclination angle is large, the rod diameter is mainly affected by the melt depth. Therefore, the actual formed rod diameter is closely related to parameters such as the thickness of the single-layer powder layer, the laser power, and the size and shape of the laser spot on the forming equipment.

[0042] Therefore, for reference Figure 1 As shown, Figure 1 This shows the existing three-dimensional objects. Figure 1 Actual objects and enlarged schematic diagrams of rods with different inclination angles.

[0043] To achieve high-precision simulation, a 3D model should be established to account for the influence of the forming geometry accuracy of the equipment. Different brands and models of SLM equipment, and even different batches of the same model, have varying laser spot parameters. This means the functional relationship between the formed rod diameter and the designed rod diameter cannot be generalized and requires fine-tuning for specific equipment. Therefore, to adapt to high-precision simulations of various equipment, the tilt angle-rod diameter function position should be reserved during the 3D modeling process. This means the function should be integrated into the modeling process as an input parameter, allowing for convenient setting of different functions for different equipment during modeling.

[0044] Therefore, for reference Figure 2 As shown, Figure 2 A flowchart illustrating an exemplary method for simulating the mechanical properties of lattice structures for additive manufacturing, based on some embodiments of this application, is shown. This application relates to a method for simulating the mechanical properties of lattice structures for additive manufacturing, comprising:

[0045] S1: Based on the 3D object image containing various diameters and support inclination angles, perform pixel point measurement to generate the forming rod diameter.

[0046] refer to Figure 3 As shown, Figure 3A schematic diagram of an exemplary specimen with various diameters and support inclination angles, representing some embodiments of this application, is shown. In practical implementation, in S1, a three-dimensional object sample is first designed using computer-aided design software (such as CAD, SolidWorks, ProE, UG, etc.). This three-dimensional object sample needs to include different diameters and support inclination angles. Then, selective laser melting (SLM), electron beam melting (EBM), or other additive manufacturing technologies are used to manufacture the three-dimensional object sample. Next, an optical microscope with a suitable magnification is selected and its pixel resolution is calibrated. The length of each pixel is recorded (the pixel length is calculated by dividing the known actual length by the corresponding number of pixels). Then, the optical microscope is used to capture images of the support parts of the three-dimensional object sample with different diameters and inclination angles to generate a three-dimensional object image containing multiple diameters and support inclination angles. Then, the image of the three-dimensional object sample is imported into image processing software (such as Python's OpenCV or ImageJ) and the image is binarized to extract edge information. Then, a pixel measurement tool is used to obtain the pixel diameter of the sample forming rod. Finally, the pixel diameter is converted into the actual diameter according to the pixel resolution of the optical microscope (the actual diameter is calculated by multiplying the pixel diameter measured in the image by the pixel length).

[0047] S2: Using the rod diameter and inclination angle as inputs and the formed rod diameter as output, a machine learning artificial neural network method is used for training to fit the inclination angle-rod diameter, and then the formed rod diameter is predicted based on the design value. The predicted value is then used to establish a lattice structure model.

[0048] refer to Figure 4 As shown, Figure 4 This diagram illustrates an exemplary conversion of a design model to a finite element simulation model based on a predicted forming rod diameter, according to some embodiments of this application. In actual implementation, the overall process in S2 is as follows:

[0049] Importing modules (libraries) for data processing, modeling, and evaluation typically includes: modules (libraries) that provide data reading and processing, modules (libraries) that provide machine learning algorithms and tools, and modules (libraries) for numerical computation. For example, in Python, libraries such as pandas, sklearn, and numpy are typically imported to support tasks such as data reading, modeling, and evaluation.

[0050] Load data containing rod diameter and tilt angle features and normalize the loaded data to bring the data to a fixed range, usually [0,1]. Then divide the processed data into training set and test set to generate a first dataset containing the training set and test set, usually with the training set accounting for 80% and the test set accounting for 20%.

[0051] When using a multilayer perceptron (MLP) to build a neural network regression model to fit the relationship between tilt angle and pole diameter, the data first needs to be preprocessed by normalization or standardization to ensure it is within a suitable range and improve the convergence speed of the neural network. Next, the structure of the neural network is defined, including the input layer, hidden layers, and output layer. The number of neurons in the hidden layers needs to be determined experimentally. Weights and biases are set through random initialization. During the forward propagation phase, the weighted sum of inputs for each neuron is calculated and then transformed using an activation function, such as ReLU or Sigmoid, to introduce nonlinearity. The loss function uses mean squared error (MSE) to calculate the difference between the predicted and actual values. During the backpropagation phase, the gradient of the loss function with respect to the weights and biases is calculated, and then the parameters are updated using the gradient descent algorithm. The calculation process involves the chain rule: the product of the gradient of the loss function with respect to the activation values, the derivative of the activation function with respect to the weighted sum, and the derivative of the weighted sum with respect to the weights. This process is repeated until a preset maximum number of iterations is reached, or the value of the loss function decreases to a preset threshold, thereby training the neural network regression model.

[0052] After each training iteration, the trained neural network regression model is used to make predictions on the test set, and the mean squared error and coefficient of determination (R²) are calculated based on the prediction results. 2 The model is evaluated using the mean squared error, which measures the average squared error between the predicted and actual values; a smaller value is better. The coefficient of determination measures the goodness of fit of the model; a value closer to 1 indicates better model performance.

[0053] Based on the design values, the diameter of the formed rod is predicted using a neural network regression model, and then the predicted values ​​are used to establish a lattice structure model.

[0054] S3: Compression tests are conducted on three-dimensional compression mechanical specimens of different diameters to obtain mechanical parameters for different diameters. Machine learning artificial neural network methods are used to train and fit the diameter-mechanical parameters to establish a mechanical test model, and then the elastic modulus of the average forming diameter is predicted.

[0055] refer to Figure 5 As shown, Figure 5This diagram illustrates an exemplary three-dimensional compression mechanical specimen compression process according to some embodiments of this application. In actual implementation, in step S3, computer-aided design software (e.g., CAD, SolidWorks, ProE, UG, etc.) is first used to design three-dimensional compression mechanical specimens with different diameters. Then, selective laser melting (SLM), electron beam melting (EBM), or other additive manufacturing technologies are used to fabricate these specimens. Next, material testing equipment (e.g., compression testing machine, universal testing machine, etc.) is used to perform mechanical compression tests on the three-dimensional compression mechanical specimens of different diameters, and the experimental results, including compressive load (force applied during compression) and displacement value (deformation of the specimen during compression), are recorded to generate stress-strain curves. Finally, the elastic modulus is calculated based on these experimental results. Where E is the elastic modulus, ΔF is the load increment per unit length, A is the cross-sectional area of ​​the three-dimensional compression mechanical specimen, Δu is the displacement increment in the loading direction, and L is the initial height of the three-dimensional compression mechanical specimen; in this way, the diameter, matching elastic modulus, and other mechanical parameters required by the experimenters for each three-dimensional compression mechanical specimen are recorded, and the corresponding mechanical data table is generated; when calculating the elastic modulus, the corresponding yield strength and compressive strength can also be calculated.

[0056] Then, this mechanics data table is used as the second dataset for machine learning, to be trained using machine learning artificial neural network methods. The specific process is as follows:

[0057] Importing modules (libraries) for data processing, modeling, and evaluation typically includes: modules (libraries) that provide data reading and processing, modules (libraries) that provide machine learning algorithms and tools, and modules (libraries) for numerical computation. For example, in Python, libraries such as pandas, sklearn, and numpy are typically imported to support tasks such as data reading, modeling, and evaluation.

[0058] Load the aforementioned mechanical data table containing diameter and elastic modulus features and normalize all the data in it to bring the data to a fixed range, usually [0,1]. Then divide the processed data into a training set and a test set to generate a second dataset containing the training set and the test set, usually with the training set accounting for 80% and the test set accounting for 20%.

[0059] A neural network regression model is constructed using a multi-sided perceptron (MLP) to fit the diameter-mechanical parameters. The average error is used as the loss function of the neural network regression model. Then, the hidden layers and the number of neurons they contain are set, and the maximum number of training iterations is set. The configured neural network regression model is then trained using backpropagation and gradient descent methods.

[0060] After each training iteration, the trained neural network regression model is used to make predictions on the test set, and the mean squared error (MSE) and coefficient of determination (R²) are calculated based on the prediction results. 2 The model is then evaluated. The formula for calculating the mean squared error is: The formula for calculating the coefficient of determination is: Where y i It is the i-th true value. This is the i-th predicted value, where n is the number of samples. It is the average of the true values.

[0061] The neural network regression model with mean square error and coefficient of determination that meet the test personnel's set values ​​is set as the mechanical test model, and then the elastic modulus of the average forming diameter is predicted based on the mechanical test model.

[0062] S4: Based on the lattice structure model and the mechanical test model, a finite element simulation model is established, and the predicted elastic modulus value is input as a material parameter into the finite element simulation model to perform mechanical property analysis. Then, the finite element simulation model is adjusted and fitted based on the generated mechanical property analysis results.

[0063] In actual implementation, in S4, the model file is first imported into the finite element software ABAQUS.

[0064] Then, material properties are set: including basic properties, plastic properties, rate-dependent properties, and damage properties, such as density, elastic modulus, Poisson's ratio, stress, strain, damage parameters, and damage evolution; in this application, the JC constitutive model is used to define the material. (T * ) m ], where σ is the equivalent flow stress, It is equivalent plastic strain. It is the equivalent plastic strain rate. It is the equivalent plastic strain rate, T m It is a dimensionless temperature, and the calculation formula is: Where T ref This is the reference temperature, T. melt A is the melting temperature, B is the initial yield stress, C is the hardening modulus, n is the hardening index, C is the strain rate hardening parameter, and m is the temperature softening index.

[0065] For example, consider setting the material density to 4.88324e-09, the elastic modulus to 125443, the Poisson's ratio to 0.342, and the material's plastic properties to 860, 450, 0.325, 1, 1650, and 1400 (where 860 is the initial yield stress, 450 is the strain corresponding to the stress, 0.325 and 1 are other plastic parameters for material hardening defined by the specific software, and 1650 and 1400 are further yield stress and strain points), and the material's rate dependence properties to 0.019 and 0.011 (where 0.019 and 0.011 are further dependent on the strain). The damage parameters of the material are set to 0.221, 32.332, 12.814, 1, 1, 0, 0, 1 (where 0.221 is the initial damage value, 32.332 and 12.814 are the damage strain or damage initiation parameters, 1 and 1 are the damage evolution control parameters, 0 and 0 are the damage influence parameters, and 1 is the final damage parameter value). The damage evolution of the material is set to 0.08*Meshsize*0.5 (where 0.08 is the damage evolution scaling factor, Meshsize is the mesh size, and 0.5 is the adjustment factor that controls the damage evolution rate and distribution).

[0066] Then, set the dynamic analysis parameters, which include analysis step size, mass scaling, target time increment, field output, and history output type.

[0067] Then, the model is assembled: the lattice structure model is meshed, corresponding material sections are assigned to each part of the structure, and the mechanical test simulation model and fixtures are assembled together according to their corresponding positional relationships to form a complete finite element simulation model.

[0068] Next, define the loads and boundary conditions for the finite element simulation model: define the external loads and boundary conditions that the structure is subjected to in the dynamic analysis, including general contact, displacement values ​​of the indenter, and complete fixed constraints at the bottom. In order to make the simulation process consistent with the actual experimental process, this part requires the experimental personnel to ensure the accuracy and rationality of the loads and boundary conditions.

[0069] Thus, the finite element simulation model is constructed, and the predicted elastic modulus value is input into the finite element simulation model as a material parameter to perform mechanical property analysis and generate mechanical property analysis results.

[0070] refer to Figures 6-12 As shown, Figure 6 A schematic diagram illustrating an exemplary mechanical simulation result of some embodiments of this application is shown. Figure 7 A schematic diagram illustrating an exemplary mechanical experiment result of some embodiments of this application is shown. Figure 8The diagram illustrates an exemplary mechanical experiment and simulation comparison of some embodiments of this application. Figure 9 The diagram illustrates a kinetic-internal energy curve of an exemplary simulation process according to some embodiments of this application. Figure 10 This illustration shows a comparison diagram of an exemplary simulation method and experimental results of some embodiments of this application. Figure 11 This paper illustrates a comparison diagram between an exemplary simulation method of this application and an unmodified simulation, representing some embodiments of this application. Figure 12 The illustration shows an exemplary simulation method of this application, experimental results, and a comparison diagram of an uncorrected simulation, representing some embodiments of this application.

[0071] Then, the generated mechanical property analysis results are compared with the actual experimental data to generate error results.

[0072] Based on the error results, error analysis is performed, and an error function is defined according to the difference between the quantified experimental values ​​and the simulated values.

[0073] For example, the error function can be defined as: Where N is the number of data points, σ sim,i The stress is the simulated value, σ exp,i The stress is the experimental value.

[0074] Based on the error function, adjust the parameters of the finite element simulation model; if the simulation results have a large error compared with the experimental results, the geometric parameters and material parameters can be adjusted.

[0075] The cyclic iteration experiment is set by the personnel to optimize the finite element simulation model to minimize the error function.

[0076] Therefore, after each optimization of the finite element simulation model, the mechanical properties are reanalyzed using the optimized finite element model.

[0077] In some embodiments of this application, reference is made to Figure 13 As shown, Figure 13 A schematic diagram of an exemplary lattice structure mechanical performance simulation system for additive manufacturing, according to some embodiments of this application, is shown. This application also relates to a lattice structure mechanical performance simulation system for additive manufacturing, comprising:

[0078] A simulation device for the mechanical properties of lattice structures for additive manufacturing using the above-mentioned simulation method.

[0079] Additive manufacturing equipment is used to form three-dimensional objects and three-dimensional compression mechanical specimens using additive manufacturing technology.

[0080] This typically includes additive manufacturing equipment such as selective laser melting (SLM) and electron beam melting (EBM), which form three-dimensional objects by depositing materials layer by layer. Its core principle is to manufacture parts layer by layer using powder or solid materials based on three-dimensional model data. It is suitable for processing complex structures, such as specimens with various diameters and column inclination angles. Additive manufacturing equipment typically includes key components such as a powder supply system, forming chamber, laser or electron beam emitter, and multi-axis platform.

[0081] A light microscope device used to capture images.

[0082] Among them, optical microscopes are devices that use optical systems to magnify and observe samples. By photographing the surface structure of the sample, they can obtain the geometric dimensions (such as diameter and tilt angle) of the additive manufacturing sample.

[0083] Materials testing equipment is used to conduct mechanical tests on materials.

[0084] Material testing equipment is an instrument used to evaluate the mechanical properties of specimens, typically including compression testing machines, tensile testing machines, fatigue testing machines, or universal testing machines. This application uses a compression testing machine to perform compression tests on additive manufacturing specimens, evaluating parameters such as elastic modulus and yield strength. Furthermore, it can clamp and load non-standard specimens (such as additively manufactured specimens), and then calculate relevant mechanical parameters using real-time acquired load-displacement curve data.

[0085] In some embodiments of this application, a simulation device for the mechanical properties of lattice structures for additive manufacturing using the above-described simulation method includes:

[0086] The image processing module is used to measure pixels based on the image of the formed 3D object containing multiple diameters and support inclination angles, and generate the forming rod diameter;

[0087] The first training module is used to take the rod diameter and inclination angle as inputs, the formed rod diameter as output, and use machine learning artificial neural network methods to train, fit the inclination angle-rod diameter, and then predict the formed rod diameter based on the design value, and use the predicted value to build a lattice structure model.

[0088] The second training module is used to conduct compression tests on three-dimensional compression mechanical specimens of different diameters, obtain mechanical parameters of different diameters, and use machine learning artificial neural network methods to train and fit diameter-mechanical parameters, establish a mechanical test model, and then predict the elastic modulus of the average forming diameter.

[0089] The finite element simulation module is used to establish a finite element simulation model based on the lattice structure model and the mechanical test model, and input the predicted elastic modulus value as a material parameter into the finite element simulation model to perform mechanical property analysis. Then, the finite element simulation model is adjusted and fitted based on the generated mechanical property analysis results.

[0090] In some embodiments of this application, reference is made to Figure 14 As shown, Figure 14 A schematic diagram of an exemplary electronic device according to some embodiments of this application is shown. The electronic device includes a memory 201 and a processor 202. The memory 201 stores a computer program that can run on the processor 202. When the processor 202 executes the computer program, it implements the methods described in the above embodiments. The number of memories 201 and processors 202 can be one or more.

[0091] The electronic device also includes:

[0092] Communication interface 203 is used to communicate with external devices and perform data exchange and transmission.

[0093] If the memory 201, processor 202, and communication interface 203 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 14 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0094] Optionally, in a specific implementation, if the memory 201, processor 202, and communication interface 203 are integrated on a single chip, then the memory 201, processor 202, and communication interface 203 can communicate with each other through an internal interface.

[0095] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor 202, implements the method provided in this application.

[0096] This application also provides a chip, which includes a processor 202 for calling and running instructions stored in a memory 201, so that a communication device equipped with the chip executes the method provided in this application.

[0097] This application also provides a chip, including: an input interface, an output interface, a processor 202 and a memory 201. The input interface, the output interface, the processor 202 and the memory 201 are connected through an internal connection path. The processor 202 is used to execute code in the memory 201. When the code is executed, the processor 202 is used to execute the method provided in the application embodiment.

[0098] It should be understood that the processor 202 mentioned above can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor 202 can be a processor that supports Advanced Reduced Instruction Set Machines (ARM) architecture.

[0099] Furthermore, the aforementioned memory 201 may include read-only memory and random access memory, and may also include non-volatile random access memory. The memory 201 may be volatile memory or non-volatile memory, or may include both. The non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory may include random access memory (RAM), which serves as an external cache. Many forms of RAM are available by way of example, but not limitation. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0100] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0101] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A simulation method for the mechanical properties of lattice structures for additive manufacturing, characterized in that, The method includes: Based on the 3D object image containing various diameters and support inclination angles, pixel point measurements are performed to generate the shaped rod diameter; The rod diameter and inclination angle are used as inputs, the formed rod diameter is used as output, and machine learning artificial neural network methods are used for training to fit the inclination angle-rod diameter, and then the formed rod diameter is predicted based on the design value, and the predicted value is used to build a lattice structure model. Compression tests were conducted on three-dimensional compression mechanical specimens of different diameters to obtain mechanical parameters for different diameters. Machine learning artificial neural network methods were used to train and fit the diameter-mechanical parameters to establish a mechanical test model, and then the elastic modulus of the average forming diameter was predicted. Based on the lattice structure model and the mechanical testing model, a finite element simulation model is established, and the predicted elastic modulus value is input into the finite element simulation model as a material parameter for mechanical property analysis. Then, the finite element simulation model is adjusted and fitted based on the generated mechanical property analysis results. The method for obtaining mechanical parameters for different diameters includes: Three-dimensional compressive mechanical specimens of different diameters were formed using additive manufacturing technology; Mechanical compression tests were conducted on three-dimensional compression mechanical specimens of different diameters using material testing equipment, and the experimental results, including compression load and displacement values, were recorded. The elastic modulus is calculated based on the experimental results, and mechanical data is generated based on the diameter of each three-dimensional compression mechanical specimen and the matched elastic modulus. The method for establishing a finite element simulation model includes: Set material properties; Set the dynamic analysis parameters, which include analysis step size, mass scaling, target time increment, field output, and history output type; The lattice structure model is meshed and assembled with a mechanical test model to form a finite element simulation model. Define the external loads and boundary conditions that the structure experiences in the dynamic analysis.

2. The simulation method according to claim 1, characterized in that, The method for generating the shaped rod diameter includes: Using additive manufacturing technology to form three-dimensional objects with various diameters and pillar angles; The three-dimensional object is photographed using a light microscope device to generate a three-dimensional object image containing various diameters and pillar tilt angles; The three-dimensional object image is binarized to obtain the pixel diameter of the forming rod; Based on the pixel resolution of the optical mirror device, the pixel diameter is converted into the actual diameter to generate the forming rod diameter.

3. The simulation method according to claim 2, characterized in that, When training using machine learning artificial neural networks, the method includes: Import the set of modules for data processing, modeling, and evaluation; Load data containing bar diameter and tilt angle features and normalize the data to generate a first dataset containing a training set and a test set; A neural network regression model is constructed using a multilayer perceptron to fit the tilt angle-bar diameter, and the neural network parameters and number of iterations are set. The neural network regression model is then trained using a training set. The trained neural network regression model is evaluated by using a test set to make predictions and calculate the mean squared error and coefficient of determination. Based on the design values, the diameter of the formed rod is predicted using a neural network regression model, and then the predicted values ​​are used to establish a lattice structure model.

4. The simulation method according to claim 1, characterized in that, When training using machine learning artificial neural networks, the method includes: Import the set of modules for data processing, modeling, and evaluation; Load data containing diameter and elastic modulus features and normalize the data to generate a second dataset containing a training set and a test set; A neural network regression model is constructed using a multilayer perceptron to fit the diameter-mechanical parameters. The neural network parameters and number of iterations are set, and then the neural network regression model is trained using a training set. The trained neural network regression model is evaluated by using a test set to make predictions and calculate the mean squared error and coefficient of determination. The neural network regression model that meets the preset requirements is set as the mechanical test model, and then the elastic modulus of the average forming diameter is predicted based on the mechanical test model.

5. The simulation method according to claim 3 or 4, characterized in that, When importing libraries for data processing, modeling, and evaluation, the method includes: Import the module set that provides data reading and data processing capabilities; Import the set of modules that provide machine learning algorithms and tools; Import the module set used for numerical computation.

6. The simulation method according to claim 3 or 4, characterized in that, When training the neural network regression model using a training set, the method includes: The average error is used as the loss function of the neural network regression model; Set the hidden layer and the number of neurons it contains, and set the maximum number of training iterations; Optimize model parameters using backpropagation and gradient descent methods.

7. The simulation method according to claim 1, characterized in that, When setting material properties, the method includes: Set the basic properties of the material, including material density, elastic modulus and Poisson's ratio; Set the plastic properties of the material, including stress and strain; Set the rate-dependent properties of the material; Set material damage properties, including damage parameters and damage evolution.

8. The simulation method according to claim 1 or 7, characterized in that, When adjusting and fitting the finite element simulation model based on the generated mechanical property analysis results, the method includes: The generated mechanical property analysis results are compared with the actual experimental data to generate error results; Based on the error results, error analysis is performed, and an error function is defined according to the difference between the quantified experimental values ​​and the simulated values. The parameters of the finite element simulation model are adjusted according to the error function. The finite element simulation model is optimized through iterative iterations to minimize the error function.

9. A simulation device for the mechanical properties of lattice structures for additive manufacturing using the simulation method described in any one of claims 1-8, characterized in that, include: The image processing module is used to measure pixels based on the image of the formed 3D object containing multiple diameters and support inclination angles, and generate the forming rod diameter; The first training module is used to take the rod diameter and inclination angle as inputs, the formed rod diameter as output, and use machine learning artificial neural network methods to train, fit the inclination angle-rod diameter, and then predict the formed rod diameter based on the design value, and use the predicted value to build a lattice structure model. The second training module is used to conduct compression tests on three-dimensional compression mechanical specimens of different diameters, obtain mechanical parameters of different diameters, and use machine learning artificial neural network methods to train and fit diameter-mechanical parameters, establish a mechanical test model, and then predict the elastic modulus of the average forming diameter. The finite element simulation module is used to establish a finite element simulation model based on the lattice structure model and the mechanical test model, and input the predicted elastic modulus value as a material parameter into the finite element simulation model to perform mechanical property analysis. Then, the finite element simulation model is adjusted and fitted based on the generated mechanical property analysis results.

10. A simulation system for the mechanical properties of lattice structures for additive manufacturing, including the simulation device described in claim 9, characterized in that, Also includes: Additive manufacturing equipment is used to form three-dimensional objects and three-dimensional compression mechanical specimens using additive manufacturing technology; Optical microscope equipment, used for capturing images; Materials testing equipment is used to conduct mechanical tests on materials.

Images