Cross-scale reverse design method and device for porous structure

Abstract

The invention discloses a cross-scale reverse design method and device for a porous structure. The method comprises the following steps: acquiring mechanical properties of the porous structure; performing distribution optimization and continuous coding on the mechanical property to obtain a mechanical property vector of the porous structure; converting the mechanical property vector into a condition control vector; and generating random Gaussian noise, inputting the random Gaussian noise and the condition control vector into a trained reverse generation model, and reversely generating a porous structure. According to the method, the accuracy of the cross-scale reverse design method is improved.

Description

Technical Field

[0001] This specification relates to the field of material structure design, and particularly to a method and device for cross-scale inverse design of porous structures. Background Art

[0002] Porous structures are material structures applicable to fields such as aerospace and medicine. Through artificial design, porous materials can meet specific requirements, such as retaining the required stiffness and strength while minimizing the mass of the workpiece as much as possible. Current porous materials are usually realized by repeating porous units (generally cube units). Researchers design the geometric structure within the unit to meet the requirements of mechanical properties. Some researchers have adjusted the dimensions of the workpiece at the stress concentration site by analyzing the stress distribution to accurately improve the mechanical properties of the workpiece. However, such optimization methods rely on finite element analysis to obtain the stress distribution, resulting in a large amount of calculation.

[0003] In recent years, with the development of machine learning and neural network technologies, data-driven inverse design methods have been developed. The data-driven neural network model establishes an inverse mapping that outputs a porous structure based on the target properties through a large amount of "porous structure - mechanical property vector" data, realizing the customized generation of porous structures. Specifically, the target properties input into the neural network serve as the control conditions for the generation network to generate a porous structure that meets the target properties. Neural network frameworks based on generative adversarial models, autoencoders, or diffusion models can all achieve this function. However, the current neural network-based inverse design solutions for porous structures still have problems with poor effects, especially for scenarios where the order of magnitude distribution of the target properties is relatively broad (cross-scale, such as 10 -5 ~10 -1 ), and it is difficult for the inverse generation model to achieve accurate customized design for all target properties within the overall range. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for cross-scale inverse design of porous structures to improve the problem of poor effects of inverse design in the cross-scale target property domain in the prior art.

[0005] This specification adopts the following technical solutions: This specification provides a method for cross-scale inverse design of porous structures, including: Obtaining the mechanical properties of the porous structure; Performing distribution optimization on the mechanical properties to obtain optimized mechanical properties; Performing continuous encoding on the optimized mechanical properties to obtain a mechanical property vector; Converting the mechanical property vector into a conditional control vector; and Generate random Gaussian noise of a porous structure, and input the random Gaussian noise and the conditional control vector into a trained inverse generation model to inversely generate a porous structure.

[0006] Optionally, the distribution optimization of the mechanical properties to obtain the optimized mechanical properties includes: Perform a linear function - logarithmic function transformation on the mechanical properties to obtain the optimized mechanical properties, where the linear function - logarithmic function transformation formula is , where the coefficient and are preset, represents the mechanical property, represents the optimized mechanical property.

[0007] Optionally, the continuous encoding of the optimized mechanical properties to obtain a mechanical property vector includes: Perform Gaussian encoding and discrete sampling on the optimized mechanical properties to obtain a mechanical property vector.

[0008] Optionally, training the inverse generation model includes: Construct a data set composed of porous structures and corresponding mechanical property vectors; Convert the mechanical property vector into a conditional control vector; Generate random Gaussian noise of the porous structure, input the random Gaussian noise and the conditional control vector into the inverse generation model, and output the generated porous structure; Calculate the difference between the generated porous structure and the porous structure in the data set as the loss function; Based on the loss function, implement the training of the inverse generation model.

[0009] Optionally, the construction of the data set composed of porous structures and corresponding mechanical property vectors includes: Perform a linear function - logarithmic function transformation on the mechanical properties of the porous structure to obtain the optimized mechanical properties, where the linear function - logarithmic function transformation formula is , where the coefficient and are preset, represents the mechanical property, represents the optimized mechanical property; Continuously encode the optimized mechanical properties to obtain a mechanical property vector; Obtain a data set composed of porous structures and corresponding mechanical property vectors.

[0010] Optionally, the continuous encoding of the optimized mechanical properties to obtain a mechanical property vector includes: Perform Gaussian encoding and discrete sampling on the optimized mechanical properties to obtain a mechanical property vector.

[0011] Optionally, the trained inverse generation model obtains a predicted noise value according to the random Gaussian noise and the conditional control vector, according to the formula Inverse generate a porous structure, where represents the random Gaussian noise, represents the conditional control vector, represents the predicted noise value, represents the inversely generated porous structure, and represents the weight value.

[0012] This specification provides a porous structure cross-scale inverse design device, including: A data acquisition module for acquiring the mechanical properties of a porous structure; A property distribution optimization module for optimizing the distribution of the mechanical properties to obtain optimized mechanical properties; A property continuous encoding module for continuously encoding the optimized mechanical properties to obtain a mechanical property vector; A property vector mapping module for converting the mechanical property vector into a conditional control vector; and An inverse generation module for generating random Gaussian noise of a porous structure and inputting the random Gaussian noise and the conditional control vector into a trained inverse generation model to inversely generate a porous structure.

[0013] This specification provides a computer-readable storage medium, including one or more programs executed by one or more processors of an electronic device, the one or more programs including instructions for executing the porous structure cross-scale inverse design method.

[0014] This specification provides an electronic device, including: one or more processors and a memory, the memory storing one or more programs, the one or more programs including instructions for executing the porous structure cross-scale inverse design method.

[0015] The above at least one technical solution adopted in this specification can achieve the following beneficial effects: Through distribution optimization, the difference in the numerical distribution at the micro scale can be enhanced, thereby improving the discrimination of control conditions at the micro scale and facilitating accurate generation; through continuous encoding optimization, simple numerical conditions are converted into vector conditions, thereby increasing the complexity and diversity of control conditions. These two optimizations can obtain high-quality conditional control vectors, thus significantly improving the accuracy of cross-scale inverse design; Since high-quality conditional control vectors are obtained through the above two optimizations, it does not occupy the training cost and inference cost of the inverse generation model, and does not incur additional computational resource and time cost consumption. Description of the Drawings

[0016] The drawings described herein are provided to further understand the present specification and form a part of the present specification. The illustrative embodiments of the present specification and their descriptions are used to explain the present specification and do not constitute an improper limitation of the present specification.

[0017] In the drawings: Figure 1 A flowchart showing the process of the multi-scale inverse design method for a porous structure in an embodiment is shown; Figure 2 An example diagram showing the inverse design effect of the multi-scale inverse design method for a porous structure in an embodiment is shown; Figure 3 The effect of optimizing the distribution of the mechanical properties in the multi-scale inverse design method for a porous structure in an embodiment is shown; Figure 4 The effect after Gaussian coding and discrete sampling of the optimized mechanical properties in the multi-scale inverse design method for a porous structure in an embodiment is shown; Figure 5 A flowchart showing the process of training the inverse generation model in the multi-scale inverse design method for a porous structure in an embodiment is shown; Figure 6 A schematic diagram showing the structure of the multi-scale inverse design device for a porous structure in an embodiment is shown; Figure 7 A schematic diagram showing the structure of an electronic device in an embodiment is shown. Detailed Description of the Embodiments

[0018] To make the purpose, technical solutions, and advantages of this specification clearer, the technical solutions of this specification will be clearly and completely described below in conjunction with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this specification without creative efforts belong to the scope of protection of this specification.

[0019] The following will detail the technical solutions provided by each embodiment of this specification in conjunction with the drawings.

[0020] Figure 1 It is a flowchart showing the process of the multi-scale inverse design method for a porous structure provided by an embodiment of the present invention. As Figure 1 shown, the multi-scale inverse design method for a porous structure in the embodiment of the present invention may include the following steps: S100: Obtain the mechanical properties of the porous structure.

[0021] Obtain the mechanical properties of a preset target (porous structure). The porous structure is expressed in voxel form. A porous structure is represented as an array of shape , where each element value is 0 or 1. 0 indicates no filling at that position, and 1 indicates filling. The porous structure 's mechanical properties , that is, the stiffness matrix, can be represented, for example, as , which contains three elements: , , .

[0022] S200: Perform distribution optimization on the mechanical properties to obtain optimized mechanical properties.

[0023] Since the mechanical properties , for example have numerical values at different scales, for numerical values at a microscopic scale, extremely small differences will also bring relatively large errors. Therefore, perform a linear function - logarithmic function transformation on the mechanical properties , for example , to widen the numerical differences of the properties at the microscopic scale. In an embodiment of this specification, performing distribution optimization on the mechanical properties to obtain optimized mechanical properties includes: performing a linear function - logarithmic function transformation on the mechanical properties to obtain optimized mechanical properties, where the linear function - logarithmic function transformation formula is where the coefficients and are preset artificially, represents the mechanical properties. For example, , or , represents the optimized mechanical properties. The optimized mechanical properties , for example basically conform to the standard Gaussian distribution.

[0024] S300: Continuously encode the optimized mechanical properties to obtain a mechanical property vector.

[0025] Convert the optimized mechanical properties , for example , into a coded vector such as through continuous encoding. In an embodiment of this specification, continuously encoding the optimized mechanical properties to obtain a mechanical property vector includes performing Gaussian encoding and discrete sampling on the optimized mechanical properties to obtain a mechanical property vector. Specifically, asFigure 4 As shown, first, take the optimized mechanical properties , for example as the central value of the Gaussian function to obtain Gaussian functions with different central values and the same variance. Next, perform discretized sampling on the Gaussian function to obtain the corresponding mechanical property vector.

[0026] S400: Convert the mechanical property vector into a conditional control vector.

[0027] Perform a concatenation operation on the mechanical property vector, for example , and convert a vector of, for example, the original shape into a shape vector. Next, input it into a property vector mapping module composed of a multi-layer fully connected network to be converted into a conditional control vector with a dimension of .

[0028] S500: Generate random Gaussian noise of the porous structure, and input the random Gaussian noise and the conditional control vector into the trained inverse generation model to inversely generate the porous structure.

[0029] Specifically, generate random Gaussian noise of the porous structure with a shape of , and input the random noise and the conditional control vector into the inverse generation model composed of a conditional diffusion model to output the generated structure . This step is a general technique used in the conditional diffusion model. The specific operations of the conditional diffusion model are divided into a diffusion process and a denoising process. In the diffusion process, the random noise is weighted and summed with the porous structure to obtain the input noise signal , , and represent weight values. In the denoising process, the conditional diffusion model predicts the original random noise according to the noise signal and the conditional control vector , and the predicted value is . Furthermore, the generated porous structure is calculated.

[0030] In an embodiment of the present specification, as Figure 5 shown, training the inverse generation model includes the following steps: S501: Construct a data set composed of porous structures and their corresponding mechanical property vectors.

[0031] For one training step, prepare a batch (with a quantity of n) of porous structures and their corresponding mechanical property vectors, for example , the shapes of the porous structure and the mechanical properties are respectively and .

[0032] S502: Convert the mechanical property vector into a conditional control vector.

[0033] Perform a concatenation operation on the mechanical property vector such as to convert a vector of the original shape such as into a shape vector. Next, input it into a property vector mapping module composed of a multi-layer fully connected network and convert it into a conditional control vector with a dimension of .

[0034] S503: Generate random Gaussian noise for the porous structure, input the random Gaussian noise and the conditional control vector into an inverse generation model, and output the generated porous structure.

[0035] Specifically, generate random Gaussian noise for a porous structure with a shape of , input the random noise and the conditional control vector into an inverse generation model composed of a conditional diffusion model, and output the generated structure .

[0036] S504: Calculate the difference between the generated porous structure and the porous structure in the dataset as a loss function.

[0037] Calculate the difference between the generated structure and the porous structure in the dataset as a loss function.

[0038] S505: Train the inverse generation model based on the loss function.

[0039] Update the network parameters in the mapping network composed of a multi-layer fully connected network and the inverse generation model by the gradient descent method; use different batches of datasets composed of porous structures and corresponding mechanical property vectors for repeated training until the network converges.

[0040] The above is the implementation of the cross-scale inverse design method for porous structures in this specification. Based on the same idea, this specification also provides a corresponding cross-scale inverse design device for porous structures, as shown in Figure 6 . The cross-scale inverse design device for porous structures in the embodiments of the present invention may include: A data acquisition module 601, configured to acquire the mechanical properties of the porous structure; ​​The property distribution optimization module 602 is used to optimize the distribution of mechanical properties to obtain optimized mechanical properties ( Figure 3 ); The property continuous encoding module 603 is used to continuously encode the optimized mechanical properties to obtain a mechanical property vector; The property vector mapping module 604 is used to transform the mechanical property vector into a conditional control vector; and The reverse generation module 605 is used to generate random Gaussian noise of the porous structure and input the random Gaussian noise and the conditional control vector into the trained reverse generation model to reversely generate the porous structure.

[0041] Figure 2 The figure shows an example diagram of the reverse design effect of the porous structure cross-scale reverse design method in the embodiment.

[0042] This specification also provides a computer-readable storage medium that stores a computer program, and the computer program can be used to execute the above Figure 1 provided porous structure cross-scale reverse design method.

[0043] This specification also provides Figure 7 a schematic structural diagram of an electronic device corresponding to Figure 1 . As Figure 7 shown, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the above Figure 1 described porous structure cross-scale reverse design method.

[0044] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to circuit structures such as diodes, transistors, switches, etc.) or software improvements (improvements to method flows). However, with the development of technology, many method flow improvements today can be regarded as direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement to a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user programming the device. Designers can program themselves to "integrate" a digital system onto a single PLD, without having to ask a chip manufacturer to design and fabricate a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compiler used in program development and writing. The original code before compilation also has to be written in a specific programming language, which is called a hardware description language (HDL), and there is not just one type of HDL, but many types, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones currently are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that by simply performing a little logical programming on the method flow using the above-mentioned several hardware description languages and programming it into the integrated circuit, it is easy to obtain the hardware circuit that implements the logical method flow.

[0045] The controller can be implemented in any suitable manner. For example, the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program code (such as software or firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of the controller include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in the form of pure computer-readable program code, it is entirely possible to logically program the method steps to enable the controller to be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, and embedded microcontrollers to achieve the same function. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be regarded as the structures within the hardware component. Or even, the devices for implementing various functions can be regarded as either software modules for implementing the method or structures within the hardware component.

[0046] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0047] For the convenience of description, when describing the above devices, they are described separately as various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0048] Those skilled in the art should understand that the embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0049] This specification is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general purpose computers, special purpose computers, embedded processors, or other programmable data processing devices to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in a block or multiple blocks.

[0050] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in a block or multiple blocks.

[0051] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the flow Figure 1 one or more flows and / or blocks Figure 1 or means for implementing the functions specified in a block or multiple blocks.

[0052] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0053] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0054] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0055] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0056] It should be understood by those skilled in the art that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0057] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0058] The various embodiments in this specification are described in a progressive manner. For the parts that are the same or similar among the various embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the method embodiments.

[0059] The above description is only for the embodiments of this specification and is not intended to limit this specification. For those skilled in the art, various changes and modifications can be made to this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this specification shall be included within the scope of the claims of this specification.

Claims

1. A cross-scale inverse design method for porous structures, characterized in that: include: Obtaining mechanical properties of porous structures; Optimizing the distribution of the mechanical properties to obtain optimized mechanical properties; Continuously encoding the optimized mechanical properties to obtain a mechanical property vector; converting the mechanical property vector into a condition control vector; as well as Generate random Gaussian noise of the porous structure, input the random Gaussian noise and the conditional control vector into the trained inverse generation model, and inversely generate the porous structure.

2. The cross-scale inverse design method of porous structure according to claim 1, characterized in that: The distribution optimization of the mechanical properties to obtain the optimized mechanical properties includes: The mechanical properties are transformed from linear function to logarithmic function to obtain optimized mechanical properties, wherein the linear function to logarithmic function transformation formula is: , where the coefficient and For pre-set, Indicates mechanical properties, Represents the optimized mechanical properties.

3. The cross-scale inverse design method of porous structure according to claim 2, characterized in that: The continuously encoding the optimized mechanical properties to obtain the mechanical property vector comprises: The optimized mechanical properties are subjected to Gaussian encoding and discrete sampling to obtain a mechanical property vector.

4. The cross-scale inverse design method of porous structure according to claim 1, characterized in that: Training the reverse generation model includes: Constructing a data set consisting of porous structures and corresponding mechanical property vectors; converting the mechanical property vector into a condition control vector; generating random Gaussian noise of the porous structure, inputting the random Gaussian noise and the conditional control vector into an inverse generation model, and outputting the generated porous structure; Calculating the difference between the generated porous structure and the porous structure in the data set as a loss function; The training of the reverse generation model is implemented based on the loss function.

5. The cross-scale inverse design method of porous structure according to claim 4, characterized in that: The constructing of a data set consisting of porous structures and corresponding mechanical property vectors includes: The mechanical properties of the porous structure are transformed from linear function to logarithmic function to obtain optimized mechanical properties, wherein the linear function to logarithmic function transformation formula is: , where the coefficient and For pre-set, Indicates mechanical properties, represents the optimized mechanical properties; Continuously encoding the optimized mechanical properties to obtain a mechanical property vector; A data set consisting of porous structures and corresponding mechanical property vectors is obtained.

6. The cross-scale inverse design method of porous structure according to claim 5, characterized in that: The continuously encoding the optimized mechanical properties to obtain the mechanical property vector comprises: The optimized mechanical properties are subjected to Gaussian encoding and discrete sampling to obtain a mechanical property vector.

7. The cross-scale inverse design method of porous structure according to claim 1, characterized in that: The trained inverse generation model obtains a predicted noise value according to the random Gaussian noise and the conditional control vector, according to the formula Reverse generation of porous structures, where represents random Gaussian noise, represents the conditional control vector, represents the predicted noise value, represents the reverse generated porous structure, and Indicates the weight value.

8. A cross-scale inverse design device for porous structures, characterized in that: include: A data acquisition module for acquiring mechanical properties of porous structures; A property distribution optimization module, used to optimize the distribution of the mechanical properties to obtain optimized mechanical properties; A property continuous encoding module, used for continuously encoding the optimized mechanical properties to obtain a mechanical property vector; A property vector mapping module, used to convert the mechanical property vector into a condition control vector; and The inverse generation module is used to generate random Gaussian noise of the porous structure and input the random Gaussian noise and the conditional control vector into the trained inverse generation model to inversely generate the porous structure.

9. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the cross-scale inverse design method of a porous structure as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: It comprises one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs comprise instructions for executing the cross-scale inverse design method of a porous structure as described in any one of claims 1-7.

Images