Device design method, device, equipment and medium
The method optimizes additive manufacturing support design using machine learning to predict and verify support structures, reducing material waste and improving efficiency and product quality.
Patent Information
- Application Number
- CN202411166467.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-12-17
AI Technical Summary
Additive manufacturing of complex geometric structures faces challenges due to insufficient support constraints, leading to deformation and interruption during the forming process, with current support design relying on experience and trial and error, resulting in inefficiency and waste.
A method using attribute data to train an agent prediction model, predict candidate support structures, filter a target result, and verify through test printing to optimize support design, employing machine learning models like support vector machines and Bayesian optimization.
This approach reduces the need for support material, saves costs, improves printing efficiency, and enhances product quality by selecting optimal support structures with fewer iterations.
Smart Images
Figure CN119150669A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of component production, and in particular to a device design method, apparatus, equipment and medium. Background Art
[0002] Additive manufacturing can produce engineering parts with complex geometric structures. Due to the overhanging structure of the workpiece, there is no support constraint of the casting or mold. Even if there is surrounding powder support, it is easy to collapse and slag during the forming process due to insufficient support force, resulting in serious deformation or even interruption of forming. Therefore, it is necessary to add support. Reasonable support is of great significance for smooth forming and ensuring the accuracy of the workpiece.
[0003] The method for determining the additive manufacturing support structure in the prior art relies on experience and trial and error, which makes the support structure determination process time-consuming and labor-intensive, and often results in incomplete support optimization, causing the part deformation to exceed the specified value, resulting in product scrapping. Summary of the invention
[0004] In view of this, the purpose of the present application is to provide a device design method, apparatus, equipment and medium to overcome the problems in the prior art.
[0005] In a first aspect, an embodiment of the present application provides a device design method, the method comprising:
[0006] Use the attribute data of existing parts to train an agent prediction model that meets the preset agent requirements;
[0007] Using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results;
[0008] Filtering out a target prediction result from the plurality of groups of candidate prediction results;
[0009] If the target prediction result passes the verification, the device to be printed is printed using the target support structure corresponding to the target prediction result to obtain the device to be printed.
[0010] In some technical solutions of the present application, the attribute data of the above-mentioned existing parts include input data and output data;
[0011] The method of using the attribute data of existing parts to train an agent prediction model that meets the preset agent requirements includes:
[0012] Using the input data to train multiple basic prediction models respectively, to obtain multiple candidate prediction models;
[0013] The proxy prediction model is selected from a plurality of candidate prediction models according to the error between the output data and the first prediction data; wherein the first prediction data is input data into the candidate prediction model, and the candidate prediction model outputs the data.
[0014] In some technical solutions of the present application, the attribute data of the existing parts include: shape data, deformation, printing parameters and support data of the supporting structure of the existing parts; the input data includes shape data, printing parameters and support data of the supporting structure of the existing parts; the output data includes deformation;
[0015] The method screens the proxy prediction model by:
[0016] From multiple candidate prediction models, the prediction model with the smallest error between the actual deformation and the predicted deformation is selected as the proxy prediction model.
[0017] In some technical solutions of the present application, the above-mentioned candidate prediction models include: support vector machine model, Gaussian process model, random forest model and other integrated learning models.
[0018] In some technical solutions of the present application, the above-mentioned screening out the target prediction result from the multiple groups of candidate prediction results includes:
[0019] Based on a statistical optimization strategy, a target prediction result is screened out from the multiple groups of candidate prediction results; wherein the statistical optimization strategy includes at least one of the following: a Bayesian optimization algorithm, an active learning algorithm, a multi-objective optimization algorithm, a transfer learning algorithm, and a genetic algorithm.
[0020] In some technical solutions of the present application, the above method verifies the target prediction result in the following manner:
[0021] Performing test printing on the device to be printed using a target support structure corresponding to the target prediction result to obtain a test printed device;
[0022] The test print device is measured, and if the measurement result meets the budgeted test requirement, the target prediction result is verified.
[0023] In some technical solutions of the present application, the above method further includes:
[0024] If the target prediction result fails to pass the verification, the attribute data of the existing parts and the newly added data are used to screen a new proxy prediction model; wherein the newly added data includes the attribute data of the device to be printed and the target prediction result, and the support data of the target support structure corresponding to the target prediction result;
[0025] The new proxy prediction model is used to make predictions again until the new target prediction results are verified.
[0026] In a second aspect, an embodiment of the present application provides a device design apparatus, the apparatus comprising:
[0027] A training module, used to train an agent prediction model that meets preset agent requirements using the attribute data of existing parts;
[0028] A prediction module, used for using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results;
[0029] A screening module, used to screen out a target prediction result from the plurality of groups of candidate prediction results;
[0030] A verification module is used to print the device to be printed using a target support structure corresponding to the target prediction result to obtain the device to be printed if the target prediction result passes the verification.
[0031] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned device design method when executing the computer program.
[0032] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned device design method are executed.
[0033] The technical solution provided by the embodiments of the present application may have the following beneficial effects:
[0034] The method of the present application includes using the attribute data of existing parts to train a proxy prediction model that meets preset proxy requirements; using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results; screening out a target prediction result from the multiple groups of candidate prediction results; if the target prediction result passes verification, printing the device to be printed using the target support structure corresponding to the target prediction result to obtain the device to be printed.
[0035] The present application can determine the target support structure using fewer times, which not only saves the amount of powder required for printing the support, but also facilitates the subsequent support removal work, saves manufacturing costs, and improves printing efficiency and product qualification rate.
[0036] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0038] Figure 1 A schematic diagram of a device design method provided in an embodiment of the present application is shown;
[0039] Figure 2 A schematic diagram of a verification process provided by an embodiment of the present application is shown;
[0040] Figure 3 A schematic diagram of a device design apparatus provided in an embodiment of the present application is shown;
[0041] Figure 4 The figure is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0042] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.
[0043] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.
[0044] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.
[0045] Additive manufacturing can produce engineering parts with complex geometric structures. Due to the overhanging structure of the workpiece, it has no support constraints from the casting or mold. Even if there is surrounding powder support, it is easy to collapse and slag during the forming process due to insufficient support force, resulting in serious deformation or even interruption of forming. Therefore, support needs to be added. Reasonable support is of great significance for smooth forming and ensuring the accuracy of the workpiece. Support design depends on many factors such as forming process, forming material, and workpiece characteristics, and is one of the process difficulties of additive manufacturing. Additional support will bring two problems. One is the additional cost and printing time, and the second is that the removal of support after printing will bring additional workload. Therefore, reasonable support design should take into account both preventing deformation and improving printability, as well as the convenience of removal.
[0046] Taking commonly used metal materials as an example, metal has high density, high forming temperature, large elastic modulus, and the workpiece is more likely to deform during the forming process. Therefore, in the process of metal additive manufacturing, not only the fixing and positioning of the support under the action of gravity is considered, but also the heat transfer and anti-deformation performance are required to be high. The current support design relies more on human experience and repeated attempts, with low design efficiency and serious waste of manpower and financial resources.
[0047] Existing additive manufacturing support optimization design methods rely on experience and trial and error, making the support optimization design time-consuming and labor-intensive. It often results in incomplete support optimization, causing part deformation to exceed the specified value, resulting in product scrapping.
[0048] Based on this, the embodiments of the present application provide a device design method, apparatus, equipment and medium, which can optimize the thickness, length and circle size of the support under a limited number of feedbacks, and finally print out parts that are not easy to deform with a smaller amount of support, which not only saves the amount of powder required for printing the support, but also facilitates the subsequent support removal work, saves manufacturing costs, and improves printing efficiency and product qualification rate. The following is a description through embodiments.
[0049] Figure 1 A schematic flow chart of a device design method provided in an embodiment of the present application is shown, wherein the method includes steps S101-S104; specifically:
[0050] S101, using the attribute data of existing parts to train an agent prediction model that meets preset agent requirements;
[0051] S102, using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results;
[0052] S103, screening out a target prediction result from the multiple groups of candidate prediction results;
[0053] S104: If the target prediction result passes the verification, the device to be printed is printed using the target support structure corresponding to the target prediction result to obtain the device to be printed.
[0054] Some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.
[0055] The technical concept of the embodiment of the present application is to quickly select a more suitable target support structure for the device to be printed. The target support structure here is used to support the device to be printed when printing, so as to ensure the integrity and accuracy of the device to be printed.
[0056] In the embodiment of the present application, when selecting a target support structure for a device to be printed, a proxy prediction model is required. The proxy prediction model is used to predict multiple candidate support structures of the device to be printed, and a more suitable target support structure is selected from the multiple groups of candidate prediction results. In order to obtain the proxy prediction model, the embodiment of the present application needs to collect attribute data of some existing parts. The initial prediction model is trained with the attribute data of these existing parts to obtain the proxy prediction model.
[0057] When using the attribute data of existing parts to train the initial prediction model, the embodiment of the present application divides the attribute data of the existing parts into input data and output data. The input data is input into the basic prediction model to obtain the first prediction data output by the basic prediction model. The basic prediction model is trained based on the first prediction data and the output data to obtain the proxy prediction model.
[0058] In the specific implementation, the existing component attribute data here includes the shape data, deformation, printing parameters and support data of the supporting structure of the existing component. Among them, the input data includes the shape data, printing parameters and support data of the supporting structure of the existing component; the output data includes the deformation. The specific data is shown in the following table:
[0059]
[0060] Specifically, the deformation of the additive component includes but is not limited to the relative deformation ratio and deformation size of the component. The geometric shape of the additive component includes but is not limited to the length, width, height and thickness of the component. The additive manufacturing printing process parameters include but are not limited to the scanning power, scanning rate, spot diameter, printing spacing and printing layer thickness. The support geometry includes but is not limited to the thickness, length, cross-sectional shape and circle size of the support.
[0061] That is to say, the embodiment of the present application inputs the shape data (length, width, thickness), deformation, printing parameters and support data of the support structure of the existing parts into the basic prediction model to obtain the predicted deformation output by the basic prediction model. Then, according to the predicted deformation and the actual deformation (the deformation of the existing parts), the basic prediction model is adjusted until the error between the predicted deformation and the actual deformation is within the preset error range, and then the proxy prediction model is obtained. During the specific training, the attribute data of the existing parts can be divided into a first part of training data and a second part of verification data.
[0062] Preferably, in order to ensure the accuracy of the proxy prediction model, the basic prediction model in the embodiment of the present application is set to be multiple, and the models obtained by training the attribute data of existing parts (referred to as candidate prediction models) are also multiple.
[0063] By screening out a proxy prediction model that meets the preset proxy requirements from multiple candidate prediction models. As mentioned above, in the embodiment of the present application, the shape data, deformation, printing parameters and support data of the support structure of the existing parts are used as input data, and the deformation is used as output data. The prediction proxy requirement here is the model with the smallest error between the actual deformation and the predicted deformation among the multiple candidate prediction models.
[0064] In the specific implementation, the prediction models to be selected include: support vector machine model, Gaussian process model, random forest model and other integrated learning models. When using these models for prediction, there are the following situations: the errors of all the prediction models to be selected are different and there are prediction models to be selected with the same errors. For the case where the errors of all the prediction models to be selected are different, they are sorted according to the size of the errors, and the prediction model to be selected with the smallest error is used as the proxy prediction model. For the case where there are prediction models to be selected with the same error, if the same error is not the smallest error, the prediction model to be selected with the smallest error is directly used as the proxy prediction model. If the same error is the smallest error, one of the prediction models to be selected with the same error can be selected as the proxy prediction model; the prediction models to be selected with the same error can also be used together as the proxy prediction model. That is to say, the proxy prediction model in the embodiment of the present application includes one model or multiple models, for example, it can be a support vector machine model or a Gaussian process model or a random forest model or other integrated learning model, or it can be any combination of a support vector machine model, a Gaussian process model, a random forest model and other integrated learning models. Specifically, the proxy prediction model is a support vector machine model; the proxy prediction model is a Gaussian process model; the proxy prediction model is a random forest model; the proxy prediction model is other integrated learning models. The proxy prediction model includes a support vector machine model and a Gaussian process model. The proxy prediction model includes a support vector machine model, a Gaussian process model, and a random forest model. The proxy prediction model includes a random forest model and other integrated learning models.
[0065] After obtaining the proxy prediction model, the proxy prediction model is used to predict the device to be printed. In actual printing work, after the device to be printed is determined, the shape data and printing parameters of the device to be printed are determined, and at the same time, multiple groups (thousands or tens of thousands) of support data of the selected support structures are determined. The embodiment of the present application is to select the target support structure for supporting the device to be printed for printing from these selected support structures. It should be noted that the device to be printed in the embodiment of the present application includes parts and features. The feature parts here are part of the parts.
[0066] By inputting the shape data, printing parameters and support data of the selected support structure of the device to be printed into the proxy prediction model, the selected prediction result output by the proxy prediction model can be obtained. For a proxy prediction model that includes only one model, the initial prediction result output by the model is directly used as the selected prediction result. For a proxy prediction model that includes multiple models, the selected prediction result is obtained by integrating the initial prediction results output by each model. The specific integration process here includes weighted calculation, fitting, calculating the average, etc.
[0067] By predicting each candidate support structure separately through the proxy prediction model, multiple groups of candidate prediction results can be obtained. At this time, the embodiment of the present application needs to screen out the target prediction result from the multiple groups of candidate prediction results.
[0068] When screening multiple groups of prediction results to be selected, the embodiment of the present application uses a statistical optimization strategy. Wherein, the statistical optimization strategy includes at least one of the following: Bayesian optimization algorithm, active learning algorithm, multi-objective optimization algorithm, transfer learning algorithm and genetic algorithm. In other words, the embodiment of the present application can use Bayesian optimization algorithm or active learning algorithm or multi-objective optimization algorithm or transfer learning algorithm or genetic algorithm alone to screen out the target prediction result from the prediction results to be selected, or it can be a combination of multiple algorithms. For example, each algorithm is used to obtain an initial prediction result. If the initial prediction results of each algorithm are consistent, the initial prediction result is directly used as the target prediction result. If there are different initial prediction results, the initial prediction result with a higher proportion can be used as the target prediction result.
[0069] After obtaining the target prediction result, the embodiment of the present application also needs to verify the target prediction result. The specific verification process is as follows: Figure 2 As shown, it includes steps S201 and S202:
[0070] S201, using a target support structure corresponding to the target prediction result to perform test printing on the device to be printed, to obtain a test printed device;
[0071] S202, measuring the test printing device, if the measurement result meets the budgeted test requirement, the target prediction result is verified.
[0072] Considering that the model is a prediction process after all, it may be different from the actual printing effect. In order to save materials, after obtaining the target prediction result, the embodiment of the present application does not directly use the target support structure to support the device to be printed for printing. Instead, it first uses the test printing method for verification. When the test printing is passed, the target support structure corresponding to the target prediction result is used to print the device to be printed to obtain the device to be printed. The test printing here can be performed using a simulation method.
[0073] If the target prediction result fails to pass the verification, it indicates that the proxy prediction model in the embodiment of the present application has not achieved the expected effect, and it is necessary to select the proxy prediction model again. In order to avoid the above situation still existing when selecting again. The embodiment of the present application retrains the basic prediction model to obtain a new prediction model to be selected. And when training again, the embodiment of the present application uses newly added data. The newly added data here includes the attribute data of the device to be printed and the target prediction result, and the support data of the target support structure corresponding to the target prediction result. That is, the embodiment of the present application allows each basic prediction model to learn the process of this failure, thereby improving the accuracy of the proxy prediction model. The specific process of training and screening the prediction model to be selected is the same as the above process, which will not be repeated here.
[0074] In an optional embodiment, the method in the present application can be implemented in the following manner: collect the geometric shape (shape data), deformation, printing process parameters and experimental data of support geometry (support data) of additive manufacturing parts (or feature parts), and establish a database for the optimization design of additive manufacturing parts support. The printing parameters are fixed, the geometric shape (length, width, height) of the parts and the experimental data of the support geometry (thickness, length, circle size) are used as input, and the deformation (%) of the parts is used as output. The data is resampled, and the input and output are fitted using a machine learning model (including a support vector machine model, a random forest model, a neural network model and other integrated learning models, etc.), and a machine learning model with small fitting error and high accuracy is selected as the proxy prediction model.
[0075] Under the same printing parameters, the components to be printed and the selected support structures are predicted and the deformation is printed out. Based on the numerical distribution of the deformation of the components predicted by the high-precision machine learning model, combined with the acquisition function in active learning (such as expected improvement), the distribution of the predicted values of the model and its uncertainty are balanced, and parameters with high expected improvement values are selected for experimental iteration. If the test results meet expectations and the measured deformation of the components meets the requirements, the iteration is stopped. If the test results do not meet expectations, the test result data is added to the training data, the model is retrained and the next round of iteration is performed until the measured deformation of the components meets the requirements and the iteration is stopped.
[0076] After two iterations of feedback, the deformation of the component was less than 0.1% under the parameters of component / component geometry length, width and height of 80mm, 50mm and 2mm respectively.
[0077] Figure 3 A schematic diagram of the structure of a device design apparatus provided in an embodiment of the present application is shown, wherein the apparatus comprises:
[0078] A training module, used to train an agent prediction model that meets preset agent requirements using the attribute data of existing parts;
[0079] A prediction module, used for using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results;
[0080] A screening module, used to screen out a target prediction result from the plurality of groups of candidate prediction results;
[0081] A verification module is used to print the device to be printed using a target support structure corresponding to the target prediction result to obtain the device to be printed if the target prediction result passes the verification.
[0082] The attribute data of the existing parts include input data and output data;
[0083] The method of using the attribute data of existing parts to train an agent prediction model that meets the preset agent requirements includes:
[0084] Using the input data to train multiple basic prediction models respectively, to obtain multiple candidate prediction models;
[0085] The proxy prediction model is selected from a plurality of candidate prediction models according to the error between the output data and the first prediction data; wherein the first prediction data is input data into the candidate prediction model, and the candidate prediction model outputs the data.
[0086] The attribute data of the existing parts include: shape data, deformation, printing parameters and support data of the supporting structure of the existing parts; the input data includes shape data, printing parameters and support data of the supporting structure of the existing parts; the output data includes deformation;
[0087] The proxy prediction models are screened by:
[0088] From multiple candidate prediction models, the prediction model with the smallest error between the actual deformation and the predicted deformation is selected as the proxy prediction model.
[0089] The candidate prediction models include: support vector machine model, Gaussian process model, random forest model and other integrated learning models.
[0090] The step of selecting a target prediction result from the plurality of groups of candidate prediction results includes:
[0091] Based on a statistical optimization strategy, a target prediction result is screened out from the multiple groups of candidate prediction results; wherein the statistical optimization strategy includes at least one of the following: a Bayesian optimization algorithm, an active learning algorithm, a multi-objective optimization algorithm, a transfer learning algorithm, and a genetic algorithm.
[0092] The target prediction results are verified by:
[0093] Performing test printing on the device to be printed using a target support structure corresponding to the target prediction result to obtain a test printed device;
[0094] The test print device is measured, and if the measurement result meets the budgeted test requirement, the target prediction result is verified.
[0095] If the target prediction result fails to pass the verification, the attribute data of the existing parts and the newly added data are used to screen a new proxy prediction model; wherein the newly added data includes the attribute data of the device to be printed and the target prediction result, and the support data of the target support structure corresponding to the target prediction result;
[0096] The new proxy prediction model is used to make predictions again until the new target prediction results are verified.
[0097] like Figure 4 As shown, an embodiment of the present application provides an electronic device for executing the device design method in the present application, the device comprising a memory, a processor, a bus, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the device design method when executing the computer program.
[0098] Specifically, the above-mentioned memory and processor may be general-purpose memory and processor, which are not specifically limited here. When the processor runs the computer program stored in the memory, the above-mentioned device design method can be executed.
[0099] Corresponding to the device design method in the present application, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned device design method are executed.
[0100] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is run, the above-mentioned device design method can be executed.
[0101] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.
[0102] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0103] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0104] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0105] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.
[0106] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims
1. A device design method, characterized in that: The method comprises: Use the attribute data of existing parts to train an agent prediction model that meets the preset agent requirements; Using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results; Filtering out a target prediction result from the plurality of groups of candidate prediction results; If the target prediction result passes the verification, the device to be printed is printed using the target support structure corresponding to the target prediction result to obtain the device to be printed.
2. The method according to claim 1, characterized in that The attribute data of the existing parts include input data and output data; The method of using the attribute data of existing parts to train an agent prediction model that meets the preset agent requirements includes: Using the input data to train multiple basic prediction models respectively, to obtain multiple candidate prediction models; The proxy prediction model is selected from a plurality of candidate prediction models according to the error between the output data and the first prediction data; wherein the first prediction data is input data into the candidate prediction model, and the candidate prediction model outputs the data.
3. The method according to claim 2, characterized in that The attribute data of the existing parts include: shape data, deformation, printing parameters and support data of the supporting structure of the existing parts; the input data include shape data, printing parameters and support data of the supporting structure of the existing parts; the output data include deformation; The method screens the proxy prediction model by: From multiple candidate prediction models, the prediction model with the smallest error between the actual deformation and the predicted deformation is selected as the proxy prediction model.
4. The method according to claim 3, characterized in that The candidate prediction models include: support vector machine model, Gaussian process model, random forest model and other integrated learning models.
5. The method according to claim 1, characterized in that: The step of selecting a target prediction result from the plurality of groups of candidate prediction results includes: Based on a statistical optimization strategy, a target prediction result is screened out from the multiple groups of candidate prediction results; wherein the statistical optimization strategy includes at least one of the following: a Bayesian optimization algorithm, an active learning algorithm, a multi-objective optimization algorithm, a transfer learning algorithm, and a genetic algorithm.
6. The method according to claim 1, characterized in that The method verifies the target prediction results in the following ways: Performing test printing on the device to be printed using a target support structure corresponding to the target prediction result to obtain a test printed device; The test printing device is measured, and if the measurement result meets the budgeted test requirement, the target prediction result is verified.
7. The method according to claim 1, characterized in that The method further comprises: If the target prediction result fails to pass the verification, the attribute data of the existing parts and the newly added data are used to screen a new proxy prediction model; wherein the newly added data includes the attribute data of the device to be printed and the target prediction result, and the support data of the target support structure corresponding to the target prediction result; The new proxy prediction model is used to make predictions again until the new target prediction results are verified.
8. A device design apparatus, characterized in that: The device comprises: A training module, used to train an agent prediction model that meets preset agent requirements using the attribute data of existing parts; A prediction module, used for using the proxy prediction model to predict the device to be printed and multiple candidate support structures of the device to be printed, respectively, to obtain multiple groups of candidate prediction results; A screening module, used to screen out a target prediction result from the plurality of groups of candidate prediction results; A verification module is used to print the device to be printed using a target support structure corresponding to the target prediction result to obtain the device to be printed if the target prediction result passes the verification.
9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the device design method as described in any one of claims 1 to 7 are performed.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the device design method according to any one of claims 1 to 7 are executed.