Integrated die casting structure topology design method and related equipment

Abstract

The invention discloses an integrated die casting structure topology design method and related equipment, and belongs to the field of integrated die castings, and the method comprises the following steps: obtaining a finite element model of an integrated die casting, carrying out characteristic decomposition, and extracting structure state parameters; constructing an integrated die casting structure topology design intelligent agent model; based on the structural state parameters, pre-training the agent model, and further completing structural topology design through the trained agent model; the agent model comprises a topological optimization strategy network for determining a structure optimization direction based on the structure state parameters and outputting optimization actions, and a structure performance evaluation network for evaluating the structure state parameters updated by the optimization actions. The structural performance evaluation network is adopted to perform structural performance evaluation on the structural state parameters of the integrated die casting, online finite element simulation in the traditional reinforcement learning process is replaced, the iteration training time can be effectively shortened, and the structural topology design optimization efficiency of the integrated die casting is improved.

Description

Technical Field

[0001] This application relates to the field of integrated die casting technology, and in particular to a method for designing the structural topology of integrated die casting and related equipment. Background Technology

[0002] Currently, integrated die casting technology greatly improves production efficiency and reduces structural weight by integrating dozens or even hundreds of traditional parts into a single die casting model.

[0003] In related technologies, the topology of a single die-cast component is typically optimized to improve its safety performance and achieve weight reduction. However, in practical applications, it has been found that traditional topology optimization methods require finite element simulation for each design iteration. For complex collision safety analyses, a single simulation can take hours or even days, resulting in a lengthy optimization cycle.

[0004] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0005] This application provides a method and related equipment for topology design of integrated die-cast parts, which can effectively shorten the iterative training time and improve the optimization efficiency of topology design for integrated die-cast parts.

[0006] On one hand, embodiments of this application provide a method for topology design of an integral die-casting component, the method comprising the following steps: A finite element model of the integral die casting is obtained, and the finite element model of the integral die casting is decomposed to extract the structural state parameters of the integral die casting. Construct an intelligent agent model for the integrated die-casting component structural topology design; Based on the structural state parameters of the integrated die casting, the intelligent agent model for the structural topology design of the integrated die casting is pre-trained, and then the structural topology design of the integrated die casting is completed through the trained intelligent agent model for the structural topology design of the integrated die casting. The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

[0007] Optionally, the integrated die-casting component structure topology design intelligent agent model further includes: an attention mechanism layer; The attention mechanism layer is used to calculate the attention state value of each node in the structure of the integral die casting; The intelligent agent model for the topology design of the integrated die-casting component calculates the attention distribution map of the integrated die-casting component structure based on the attention state value of each node and a predefined stress distribution query vector.

[0008] Optionally, the method further includes: The topology optimization strategy network of the intelligent agent model for the integrated die-casting structure determines the optimization direction of the integrated die-casting structure based on the attention distribution map of the integrated die-casting structure.

[0009] Optionally, the method further includes: The structural performance evaluation network of the integrated die-casting structure topology design intelligent agent model calculates a reward function based on the structural state parameters of the integrated die-casting after the optimized action update, and then evaluates the structural performance of the integrated die-casting. The reward function includes static performance reward, collision safety performance reward, lightweight reward, and continuity reward.

[0010] Optionally, the step of obtaining the finite element model of the integral die casting and performing feature decomposition on the finite element model of the integral die casting to extract the structural state parameters of the integral die casting includes: Obtain the finite element model of the integral die-cast part; The design domain of the finite element model of the integral die-cast part is determined, and feature decomposition is performed by graph structure discretization; Based on the graph structure discretization results, the nodes and edges of the integral die casting are extracted to form a graph structure representation of the integral die casting. The representation is then encoded by a state encoder, and the encoding result is used as the structural state parameter of the integral die casting.

[0011] Optionally, the integrated die-casting component structure topology design intelligent agent model is trained based on the following steps: Initialize the model parameters of the intelligent agent model for the topology design of the integrated die-casting component structure; By employing an experience replay and strategy update algorithm, the topology optimization strategy network and structural performance evaluation network of the integrated die-casting structure topology design intelligent agent model are updated until the integrated die-casting structure topology design intelligent agent model converges, and the trained integrated die-casting structure topology design intelligent agent model is output.

[0012] On the other hand, embodiments of this application provide an integrated die-casting part structure topology design device, the device comprising: The data acquisition module is used to acquire the finite element model of the integral die casting and to perform feature decomposition on the finite element model of the integral die casting to extract the structural state parameters of the integral die casting. The model building module is used to build an intelligent agent model for the topology design of an integrated die-casting component structure. The topology design module is used to pre-train the intelligent agent model for the topology design of the integrated die casting based on the structural state parameters of the integrated die casting, and then complete the topology design of the integrated die casting structure through the trained intelligent agent model for the topology design of the integrated die casting structure. The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

[0013] On the other hand, embodiments of this application provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0014] On the other hand, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0015] On the other hand, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0016] This application embodiment uses a structural performance evaluation network to evaluate the structural state parameters of an integral die-cast part, replacing the online finite element simulation in the traditional reinforcement learning process. This can effectively shorten the iterative training time and improve the efficiency of structural topology design optimization for integral die-cast parts. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the implementation environment for an integrated die-casting component structure topology design method provided in an embodiment of this application; Figure 2 This is a flowchart illustrating a method for topology design of an integral die-casting component provided in an embodiment of this application. Figure 3 This is a flowchart illustrating a training-integrated die-casting structural topology design intelligent agent model provided in an embodiment of this application; Figure 4 This is a schematic diagram of the framework of a training integrated die-casting part structure topology design intelligent agent model provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an integrated die-casting part topology design device provided in an embodiment of this application; Figure 6 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0019] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0020] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0022] Currently, integrated die casting technology greatly improves production efficiency and reduces structural weight by integrating dozens or even hundreds of traditional parts into a single die casting model.

[0023] In related technologies, the topology of a single die-cast component is typically optimized to improve its safety performance and achieve weight reduction. However, in practical applications, it has been found that traditional topology optimization methods require finite element simulation for each design iteration. For complex collision safety analyses, a single simulation can take hours or even days, resulting in a lengthy optimization cycle.

[0024] In view of this, this application provides a method and related equipment for topology design of integrated die-cast parts. By using a structural performance evaluation network to evaluate the structural state parameters of integrated die-cast parts, the method replaces online finite element simulation in the traditional reinforcement learning process, which can effectively shorten the iterative training time and improve the optimization efficiency of topology design of integrated die-cast parts.

[0025] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0026] The specific implementation methods of the embodiments of this application will be described in detail below with reference to the accompanying drawings. First, a method for topology design of an integral die-casting part provided in the embodiments of this application will be described with reference to the accompanying drawings.

[0027] Please refer to Figure 1 , Figure 1 This is a schematic diagram of the implementation environment for an integrated die-casting component structure topology design method provided in this application embodiment. In this implementation environment, the main hardware and software components involved include a terminal processor 110 and a server 120.

[0028] Specifically, the terminal processor 110 may be equipped with a control program for the integrated die-casting structure topology design method, and the server 120 serves as the backend server for this control program. The terminal processor 110 and the backend server 120 are connected for communication. The integrated die-casting structure topology design method provided in this embodiment can be executed on the terminal processor 110 side.

[0029] Server 120 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0030] In addition, server 120 can also be a node server in a blockchain network.

[0031] The terminal processor 110 and the server 120 can establish a communication connection via a wireless network. This wireless network uses standard communication technologies and / or protocols. The network can be the Internet or any other network, including but not limited to a Local Area Network (LAN), Metropolitan Area Network (MAN), Wide Area Network (WAN), mobile, or any combination of wireless networks, private networks, or virtual private networks. Furthermore, these hardware and software components can use the same or different communication connection methods; this application does not impose specific limitations in this regard.

[0032] Of course, this is understandable. Figure 1 The implementation environment described in this application is only one of the optional application scenarios in the integrated die-casting structure topology design method provided in this embodiment. The actual application is not fixed. Figure 1 The software and hardware environment shown is not specifically limited in this application.

[0033] like Figure 2 As shown, Figure 2 This is a flowchart illustrating a method for topology design of an integral die-casting part provided in an embodiment of this application, specifically including but not limited to steps 100 to 300.

[0034] Step 100: Obtain the finite element model of the integral die casting, and perform feature decomposition on the finite element model of the integral die casting to extract the structural state parameters of the integral die casting.

[0035] In this embodiment of the application, the finite element model of the integral die-casting part whose topology is to be optimized is first obtained. The finite element model of the integral die-casting part includes multiple integral die-casting part configurations. The configuration parameters of each integral die-casting part configuration may include data parameters such as the total length, total width, total height of the rear floor, the thickness of the longitudinal beam, the thickness of the transverse beam, the height of the reinforcing rib, the number of reinforcing ribs, and the diameter of the mounting base.

[0036] Furthermore, the finite element model of the integral die casting is subjected to feature decomposition to extract the structural state parameters of the integral die casting, which are used as input data for the subsequent structural topology design of the integral die casting.

[0037] Specifically, as an optional implementation, the process of obtaining a finite element model of the integral die-casting and performing feature decomposition on the finite element model of the integral die-casting to extract the structural state parameters of the integral die-casting includes: Obtain the finite element model of the integral die-cast part; The design domain of the finite element model of the integral die-cast part is determined, and feature decomposition is performed by graph structure discretization; Based on the graph structure discretization results, the nodes and edges of the integral die casting are extracted to form a graph structure representation of the integral die casting. The representation is then encoded by a state encoder, and the encoding result is used as the structural state parameter of the integral die casting.

[0038] In the embodiments of this application, please refer to Figure 3 , Figure 3 This is a flowchart illustrating a training process for an intelligent agent model of a single die-cast part structure topology design, provided in an embodiment of this application. First, a finite element model of the single die-cast part is selected. The side and bottom reinforcing rib distribution areas of the finite element model of the single die-cast part are defined as the design domain. At the same time, the surfaces in the design domain are discretized into a graphical structure based on finite element mesh elements.

[0039] Specifically, when discretizing the surface in the design domain into a graph structure based on finite element mesh elements, the central node of each mesh element in the design domain can be regarded as a node in the graph structure, and the connection relationship between mesh elements can be abstracted as edges in the graph. Thus, based on the extracted nodes and edges, a graph structure representation of the integral die-casting structure can be constructed. ,in It is a graph structure representation. It is a set of nodes. It is a set of edges.

[0040] In practical applications, each node in a node set can be expressed as an eigenvector after feature decomposition, specifically as a two-dimensional tensor. It mainly includes two channels of information: the first channel information, which is the spatial coordinate data of the node, and the second channel information, which is the probability of the existence of the material to which the node belongs. That is, the node with stiffeners in the design domain is set to 1, and the node without stiffeners is set to 0.

[0041] Furthermore, by using the trained graph convolutional network as a state encoder, the graph structure representation of the integral die casting is taken as the data input. The information of neighboring nodes is aggregated through multi-layer graph convolution operations, and finally the embedding vector is output for each node in the integral die casting as the structural state parameter of the integral die casting.

[0042] Step 200: Construct an intelligent agent model for the integrated die-casting component's structural topology design; The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

[0043] In this embodiment, an intelligent agent model for the topology design of an integrated die-casting part is constructed to perform optimization of the topology design of the integrated die-casting part.

[0044] Specifically, such as Figure 3 As shown, the intelligent agent model for topology design of integrated die-cast parts mainly includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network can be constructed based on algorithms such as Proximal Policy Optimization (PPO) and Q-learning, while the structural performance evaluation network can be constructed based on graph convolutional networks.

[0045] In practical applications, the topology optimization strategy network can receive the structural state parameters of the integrated die-casting part, execute a strategy update algorithm to determine the structural optimization direction of the integrated die-casting part's structural topology design, and output optimization actions. These optimization actions are then used to update the structural state parameters of the integrated die-casting part. Furthermore, the structural performance evaluation network can evaluate the structural performance of the integrated die-casting part after the optimization actions have been updated, thereby determining the optimization performance effect of the current structural optimization direction of the integrated die-casting part's structural topology design based on the evaluation results.

[0046] Specifically, as an optional implementation, the integrated die-casting structure topology design intelligent agent model further includes: an attention mechanism layer; The attention mechanism layer is used to calculate the attention state value of each node in the structure of the integral die casting; The intelligent agent model for the topology design of the integrated die-casting component calculates the attention distribution map of the integrated die-casting component structure based on the attention state value of each node and a predefined stress distribution query vector.

[0047] In the embodiments of this application, such as Figure 3 As shown, after constructing the graphical representation of the integral die casting and obtaining the structural state parameters through the state encoder, the performance response parameter values ​​of the current integral die casting can be further defined. Specifically, these can be obtained through finite element calculation, including the maximum stress of the integral die casting under the rear impact condition, the maximum displacement of the current integral die casting, the maximum intrusion during the collision process, the total energy absorbed during the collision process, the peak impact force during the collision process, and the mass of the integral die casting.

[0048] Furthermore, in the last layer of the state encoder, graph-level attention pooling is introduced to compute an attention state value for each node. , , n The number of nodes is represented by the attention score, which reflects the global importance of each node under the current mechanical state.

[0049] Furthermore, a predefined set of learnable stress distribution query vectors is used. ,in, Represents the first mechanical bearing area. This represents the second mechanical bearing zone, with the mechanical bearing capacity decreasing sequentially. That is, the mechanical bearing capacity of the second mechanical bearing zone is lower than that of the first mechanical bearing zone, and so on, until it eventually decays to the point of having no bearing capacity. This is a material redundancy area.

[0050] In practical applications, the intelligent agent model for topology design of integrated die-cast parts can obtain an attention distribution map by calculating the attention distribution between each query vector and the attention state value of each node. The elements in the attention distribution map represent the important actions of each node for mechanical load-bearing intentions. Therefore, when executing topology optimization actions and determining the optimization direction of the integrated die-cast part structure, the topology optimization strategy network of the intelligent agent model can determine the action strategies for different mechanical load-bearing regions based on the attention distribution map of the integrated die-cast part structure.

[0051] Specifically, such as Figure 3 As shown, an executable structural topology optimization action space can be predefined in the topology optimization strategy network of the intelligent agent model for integrated die-casting structure topology design. The topology optimization action space can act on the structural nodes of the integrated die-casting graph, and can realize the deletion or addition of nodes and the edges connecting nodes to other nodes. Furthermore, when updating the current structural state of the integrated die-casting through the output optimization action, the node attributes can be adjusted according to the specific action space. For example, if the optimization action is determined to be adding a node, the existence probability of the material to which the node belongs is changed from 0 to 1 in the second channel information of the attribute of the added node. Correspondingly, if it is deleting a node, the existence probability of the material to which the node belongs is changed from 1 to 0 in the second channel of the attribute of the deleted node.

[0052] Furthermore, the topology optimization strategy network can determine the optimization direction of the integrated die-casting structure based on the attention distribution map of the integrated die-casting structure. For example, for the query vector... The attention distribution map highlights the areas with the most severe stress concentration. The topology optimization strategy network can identify these areas as areas requiring priority in adding nodes for optimization. For the query vector... The attention distribution map highlights the current material redundancy area, and the topology optimization strategy network can identify it as a priority to perform node deletion action on this mechanically load-bearing area.

[0053] Therefore, this application uses an attention mechanism to guide the intelligent agent model of the integrated die-casting structure topology design to prioritize the high-stress-bearing area when performing optimization actions. This not only significantly improves training efficiency and final performance, but also enhances the practicality and reliability of the topology design results.

[0054] Specifically, as an optional implementation, the method further includes: The structural performance evaluation network of the integrated die-casting structure topology design intelligent agent model calculates a reward function based on the structural state parameters of the integrated die-casting after the optimized action update, and then evaluates the structural performance of the integrated die-casting. The reward function includes static performance reward, collision safety performance reward, lightweight reward, and continuity reward.

[0055] In this embodiment of the application, after the topology optimization strategy network of the integrated die-casting structure topology design intelligent agent model determines the optimization direction of the integrated die-casting structure and outputs the optimization action, the structural state parameters of the integrated die-casting are updated, and the structural performance of the integrated die-casting after the optimization action is updated is evaluated.

[0056] Specifically, the topology optimization strategy network can evaluate the structural performance of a monolithic die-cast part by calculating a reward function. This reward function includes static performance rewards, collision safety performance rewards, lightweighting rewards, and continuity rewards.

[0057] In practical applications, the reward function can be expressed as the following formula (1): (1) in, For the reward function; For static performance rewards; Awards for collision safety performance; Lightweight rewards; For continuous rewards; , , as well as Static performance bonuses Collision safety performance awards Lightweight rewards Continuous rewards The weighting coefficients can be dynamically adjusted during the pre-training process of the intelligent agent model for the topology design of the integrated die-casting structure.

[0058] Among them, the static performance bonus is used to evaluate the penalty for exceeding the standard stress and displacement, and can be calculated by the following formula (2): (2) in, The calculated maximum stress of the current integral die-cast part; The maximum allowable stress for a pre-set integral die-cast part; This is the calculated maximum displacement of the current integral die-cast part; The maximum allowable displacement of the integral die-cast part is set in advance.

[0059] Therefore, the static performance bonus uses a logarithmic form to strongly penalize situations exceeding allowable stress or displacement. ≤ as well as ≤ When the calculated static performance reward is 0, the calculated static performance reward will generate a penalty when the above control quantity exceeds the limit. The logarithm can make the penalty monotonically increase with the increase of the limit, but the rate of increase is controlled, providing a smoother and more saturated control effect.

[0060] Furthermore, the collision safety bonus can be calculated using the following formula (3): (3) in, This is a penalty item for intrusion volume; This is the calculated maximum intrusion amount under the collision condition; The maximum permissible intrusion amount for a pre-set integral die-cast part; Rewards for absorbing total energy; This represents the total energy absorbed under collision conditions. This is the initial kinetic energy; This is a penalty for the peak collision force. This represents the calculated peak collision force under the collision conditions. The maximum permissible peak impact force of the pre-set integral die-cast part; , , The weight coefficients for the intrusion penalty, energy absorption reward, and peak collision force penalty, respectively, can be dynamically adjusted during the pre-training process of the intelligent agent model for the integrated die-casting structure topology design.

[0061] Therefore, the collision safety reward takes into account the amount of intrusion, the proportion of energy absorbed, and the peak collision force, effectively reflecting the multi-dimensional objectives of the collision condition.

[0062] Furthermore, the lightweight reward can be calculated using the following formula (4): (4) in, The lightweight factor can be preset according to specific usage requirements. To ensure the quality of current integrated die-cast parts; The reference value for the pre-set integral die-cast part's baseline mass.

[0063] Therefore, the lightweight incentive is directly penalized by the difference between the current mass and the baseline mass, driving the one-piece die casting to reduce weight.

[0064] Furthermore, the continuous reward can be calculated using the following formula (5): (5) in, n The number of nodes in the graphical representation of a single die-cast part; Represents a node i The connectivity of a node can be determined by the number of nodes whose material existence probability to this node is 1. The target connectivity can be preset, for example, it can be set to 1.

[0065] Therefore, continuity rewards encourage continuous force transmission paths and avoid overly discretized "isolated pieces" by evaluating the degree of connectivity and penalizing the degree of connectivity against the target degree of connectivity, thereby improving manufacturability and structural integrity.

[0066] In practical applications, the reward function is calculated by calculating the structural state parameters of the integrated die casting after the optimized action update. This serves as the basis for evaluating the structural performance of the integrated die casting. By flexibly integrating various performance indicators such as collision safety, lightweighting, and force transmission path continuity, multi-objective Pareto optimal design can be achieved.

[0067] Step 300: Based on the structural state parameters of the integrated die casting, pre-train the intelligent agent model for the structural topology design of the integrated die casting, and then complete the structural topology design of the integrated die casting through the trained intelligent agent model for the structural topology design of the integrated die casting.

[0068] In this embodiment, after obtaining the structural state parameters of the integrated die casting and constructing the intelligent agent model for the structural topology design of the integrated die casting, the intelligent agent model for the structural topology design of the integrated die casting is pre-trained based on the structural state parameters of the integrated die casting as training data. Finally, the structural topology design of the integrated die casting is completed through the trained intelligent agent model for the structural topology design of the integrated die casting.

[0069] Specifically, as an optional implementation, the integrated die-casting component structure topology design intelligent agent model is trained based on the following steps: Initialize the model parameters of the intelligent agent model for the topology design of the integrated die-casting component structure; By employing an experience replay and strategy update algorithm, the topology optimization strategy network and structural performance evaluation network of the integrated die-casting structure topology design intelligent agent model are updated until the integrated die-casting structure topology design intelligent agent model converges, and the trained integrated die-casting structure topology design intelligent agent model is output.

[0070] In the embodiments of this application, such as Figure 3 As shown, when pre-training the intelligent agent model for topology design of integrated die-casting structure, the model parameters of the intelligent agent model for topology design of integrated die-casting structure can be initialized first. Then, the topology optimization strategy network and structural performance evaluation network of the intelligent agent model for topology design of integrated die-casting structure can be updated by adopting the experience playback and strategy update algorithm.

[0071] Specifically, for each training round, the graph structure representation corresponding to the topology optimization design domain of the integrated die-casting structure is initialized. The current graph structure representation is encoded using a graph convolutional network to obtain the embedding vector of each node, which serves as the structural state parameter of the integrated die-casting. Furthermore, the optimization direction of the structure is determined by the topology optimization policy network, and optimization actions are output. The material density of the corresponding node in the graph is updated according to the optimization actions, thereby obtaining the updated structural state parameters of the integrated die-casting and inputting them into the structural performance evaluation network. The reward function is calculated as the evaluation result. The initial structural state parameters, the updated structural state parameters, the optimization actions, and the reward function are used to construct an experience tuple and store it in the experience replay buffer. Experience data is sampled from the experience buffer, and the topology optimization policy network and the structural performance evaluation network are updated through the policy update algorithm.

[0072] Specifically, please refer to Figure 4 , Figure 4This is a schematic diagram of the framework for training an intelligent agent model for the topology design of an integrated die-casting component, as provided in an embodiment of this application. A graph representation of the integrated die-casting component is read from the experience replay buffer. The structural state parameters of the integrated die-casting component are determined through graph convolutional network encoding and input into the topology optimization strategy network. The topology optimization strategy network performs data dimensionality upscaling through matrix operations and convolutional operations through multiple convolutional blocks, combined with the residual mechanism set between each convolutional block, to finally determine the optimization direction of the structure, output optimization actions, and update the structural state of the integrated die-casting component. The updated structural state parameters of the integrated die-casting component are then input into the structural performance evaluation network. The structural performance evaluation network performs data processing operations such as dimensionality upscaling and dimensionality reduction on the structural state parameters through multiple matrix operations, and finally calculates the reward function as the evaluation result. Thus, an experience tuple is constructed using the initial structural state parameters, the updated structural state parameters, the optimization actions, and the reward function and stored in the experience replay buffer, completing the current training round.

[0073] Understandable, Figure 4 The number of circles shown in the topology optimization strategy network and structural performance evaluation network is only used to illustrate the process of data dimension change and does not represent the specific number of data dimensions. This application does not impose a specific limit on the number of data dimensions, which can be preset according to the specific scenario requirements.

[0074] In practical applications, the topology optimization policy network and the structural performance evaluation network can be updated by selecting policy update algorithms such as Proximal Policy Optimization (PPO) or Soft Actor-Critic (SAC).

[0075] Furthermore, by repeating the experience replay and policy update algorithm for each round mentioned above, until the intelligent agent model for the topology design of the integrated die-casting part converges, the pre-training process of the intelligent agent model for the topology design of the integrated die-casting part ends.

[0076] Finally, by outputting the trained intelligent agent model of the integrated die-casting structure topology design, the optimization design of the integrated die-casting structure topology is completed.

[0077] Therefore, this application uses a structural performance evaluation network to evaluate the structural state parameters of the integrated die casting, replacing the online finite element simulation in the traditional reinforcement learning process. This can effectively shorten the iterative training time and improve the efficiency of structural topology design optimization for integrated die castings.

[0078] Please see Figure 5 , Figure 5 This is a schematic diagram of a topology design device for an integrated die-casting part provided in an embodiment of this application. This application also provides a topology design device for an integrated die-casting part, which can implement the above-mentioned topology design method for an integrated die-casting part. The device includes: The data acquisition module 510 is used to acquire the finite element model of the integral die casting and to perform feature decomposition on the finite element model of the integral die casting to extract the structural state parameters of the integral die casting. Model building module 520 is used to build an intelligent agent model for the topology design of an integrated die-casting component structure. The topology design module 530 is used to pre-train the intelligent agent model for topology design of the integrated die casting based on the structural state parameters of the integrated die casting, and then complete the topology design of the integrated die casting structure through the trained intelligent agent model for topology design of the integrated die casting structure. The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

[0079] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0080] Please see Figure 6 , Figure 6 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. The electronic device includes: The processor 601 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 602 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 602 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 602 and is called and executed by the processor 601 using the methods described in the embodiments of this application. The input / output interface 603 is used to implement information input and output; The communication interface 604 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 605 transmits information between various components of the device (e.g., processor 601, memory 602, input / output interface 603, and communication interface 604); The processor 601, memory 602, input / output interface 603, and communication interface 604 are connected to each other within the device via bus 605.

[0081] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0082] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0083] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0084] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0085] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0086] This application provides a method and related equipment for topology design of integrated die-cast parts. It uses a structural performance evaluation network to evaluate the structural state parameters of integrated die-cast parts, replacing online finite element simulation in the traditional reinforcement learning process. This can effectively shorten the iterative training time and improve the optimization efficiency of topology design for integrated die-cast parts.

[0087] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0088] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0089] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0090] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0091] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0092] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0093] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0094] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0095] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for topology design of an integral die-casting component, characterized in that, The method includes the following steps: A finite element model of the integral die casting is obtained, and the finite element model of the integral die casting is decomposed to extract the structural state parameters of the integral die casting. Construct an intelligent agent model for the integrated die-casting component structural topology design; Based on the structural state parameters of the integrated die casting, the intelligent agent model for the structural topology design of the integrated die casting is pre-trained, and then the structural topology design of the integrated die casting is completed through the trained intelligent agent model for the structural topology design of the integrated die casting. The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

2. The method according to claim 1, characterized in that, The integrated die-casting component structure topology design intelligent agent model also includes: an attention mechanism layer; The attention mechanism layer is used to calculate the attention state value of each node in the structure of the integral die casting; The intelligent agent model for the topology design of the integrated die-casting component calculates the attention distribution map of the integrated die-casting component structure based on the attention state value of each node and a predefined stress distribution query vector.

3. The method according to claim 2, characterized in that, The method further includes: The structural performance evaluation network of the integrated die-casting component structure topology design intelligent agent model determines the optimization direction of the integrated die-casting component structure based on the attention distribution map of the integrated die-casting component structure.

4. The method according to claim 1, characterized in that, The method further includes: The topology optimization strategy network of the intelligent agent model for the integrated die-casting structure topology design calculates a reward function based on the structural state parameters of the integrated die-casting after the optimization action update, and then evaluates the structural performance of the integrated die-casting. The reward function includes static performance reward, collision safety performance reward, lightweight reward, and continuity reward.

5. The method according to claim 1, characterized in that, The process of obtaining a finite element model of the integral die-casting and performing feature decomposition on the finite element model to extract the structural state parameters of the integral die-casting includes: Obtain the finite element model of the integral die-cast part; The design domain of the finite element model of the integral die-cast part is determined, and feature decomposition is performed by graph structure discretization; Based on the graph structure discretization results, the nodes and edges of the integral die casting are extracted to form a graph structure representation of the integral die casting. The representation is then encoded by a state encoder, and the encoding result is used as the structural state parameter of the integral die casting.

6. The method according to claim 1, characterized in that, The intelligent agent model for the integrated die-casting component structure topology design was trained based on the following steps: Initialize the model parameters of the intelligent agent model for the topology design of the integrated die-casting component structure; By employing an experience replay and strategy update algorithm, the topology optimization strategy network and structural performance evaluation network of the integrated die-casting structure topology design intelligent agent model are updated until the integrated die-casting structure topology design intelligent agent model converges, and the trained integrated die-casting structure topology design intelligent agent model is output.

7. A topology design device for an integrated die-casting component, characterized in that, The device includes: The data acquisition module is used to acquire the finite element model of the integral die casting and perform feature decomposition on the finite element model of the integral die casting to extract the structural state parameters of the integral die casting. The model building module is used to build an intelligent agent model for the topology design of an integrated die-casting component structure. The topology design module is used to pre-train the intelligent agent model for the topology design of the integrated die casting based on the structural state parameters of the integrated die casting, and then complete the topology design of the integrated die casting structure through the trained intelligent agent model for the topology design of the integrated die casting structure. The integrated die-casting component structure topology design intelligent agent model includes a topology optimization strategy network and a structural performance evaluation network. The topology optimization strategy network determines the structural optimization direction of the integrated die-casting component based on the structural state parameters of the integrated die-casting component and outputs optimization actions. The structural performance evaluation network evaluates the structural performance of the integrated die-casting component after the structural state parameters are updated by the optimization actions.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

Images