Method, device, equipment and storage medium for determining interface between multiple substances

Through the high-precision format solution based on the initialization parameters of multi-matter and the time-propulsion step length, combined with the physical information neural network method, the problems of low calculation accuracy and large numerical dissipation in the interface determination method between multi-matters are solved, and the multi-matter is realized Precise simulation of interactions.

CN118762763BActive Publication Date: 2024-11-12国家超级计算天津中心 +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411239368.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-05
Publication Date
2024-11-12
Estimated Expiration
2044-09-05

AI Technical Summary

Technical Problem

The existing methods for determining the interface between multiple substances have low calculation accuracy and high numerical dissipation, which is difficult to meet the needs of accurate simulation of interactions between multiple substances.

Method used

Using the initialization parameters based on multi-matter, determine the spatial discrete solution result of the first precision format and the time-progress solution result of the second precision format. Combined with the physical information neural network method, solve the interface control model of multi-matter and obtain the volume fraction of multi-matter. Distribute and obtain interface data through multi-layer grid cell state processing.

Benefits of technology

The accurate and efficient simulation of multi-matter interaction problems are achieved, the accuracy of interface data is improved, and the problems of low calculation accuracy and large numerical dissipation in the prior art are solved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118762763B_ABST
    Figure CN118762763B_ABST
Patent Text Reader

Abstract

The present disclosure relates to a method, device, equipment and storage medium for determining interfaces between multiple substances. Based on the initialization parameters of multiple substances and the target time advancement step, determine the spatial discrete solution result in the first precision format and the time advancement solution result in the second precision format; input the spatial discrete solution result and the time advancement solution result into the interface control model of multiple substances, and solve them using the physical information neural network method to determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment; perform grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of multiple substances, and obtain multi-layer grid unit state data on both sides of the multi-substance interface; if the current moment is greater than the preset end time, output the state data as the interface data between multiple substances. Therefore, combined with the PINNs‑VOF interface capture method, accurate and efficient simulation of multi-substance interaction problems can be achieved, and problems related to the field of multi-substance simulation can be solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the technical field of multi-material simulation, and in particular to a method, device, equipment and storage medium for determining interfaces between multiple materials. Background Art

[0002] For the simulation of multi-material interaction problems, it is crucial to accurately capture the interface between different materials. The current mainstream methods for determining the interface between multiple materials include the multi-material interface control method (Volume of Fluid, VOF) and the numerical (Level set) method for simulating the shape and evolution of objects.

[0003] Specifically, the VOF method in the prior art uses the material volume fraction to implicitly express the interface, and the Level-Set method uses the signed distance to characterize the interface. In practical applications, the VOF method involves relatively complex geometric operations, complex logic, large amount of calculation, and is not easy to handle the determination of interfaces between three or more materials; the Level-Set method directly uses a high-precision numerical format to solve the signed distance function, which has a large conservation problem, and in order to avoid the problem of transitional diffusion and transitional contraction of the signed distance function, an initialization operation is required.

[0004] It can be seen that the methods for determining interfaces between multiple substances in the prior art have low calculation accuracy, large numerical dissipation, and are difficult to meet the needs of accurate simulation of interactions between multiple substances. Summary of the invention

[0005] In order to solve the above technical problems, the present disclosure provides a method, device, equipment and storage medium for determining interfaces between multiple substances.

[0006] In a first aspect, the present disclosure provides a method for determining an interface between multiple substances, the method comprising:

[0007] Determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and determine a time marching solution result in a second precision format based on the initialization parameters of the multi-substances and a target time marching step size;

[0008] Inputting the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of multiple materials, solving the interface control model of multiple materials by using the physical information neural network method, and determining the volume fraction distribution of multiple materials after the target time advancement step at the current moment;

[0009] Performing grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances to obtain multi-layer grid unit state data on both sides of the multi-substance interface;

[0010] If the current time is greater than the preset end time, the state data of the multi-layer grid units on both sides of the multi-substance interface are used as the interface data between the multi-substances.

[0011] In some embodiments of the present disclosure, determining the spatial discretization solution result in the first precision format based on the initialization parameters of multiple substances includes:

[0012] A third-order WENO format is used to perform spatial discretization solution on the initialization parameters of the multi-material, and a solution result in the third-order WENO format is obtained as a spatial discretization solution result in the first precision format.

[0013] In some embodiments of the present disclosure, determining the time-marching solution result in the second precision format based on the initialization parameters of the multi-substance and the target time-marching step size includes:

[0014] The second-order RK format is adopted to perform a time-marching solution on the initialization parameters according to the target time-marching step, and a solution result in the second-order RK format is obtained as the time-marching solution result in the second precision format.

[0015] In some embodiments of the present disclosure, the step of inputting the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of multiple substances, solving the interface control model of multiple substances by using a physical information neural network method, and determining the volume fraction distribution of multiple substances after the target time advancement step at the current moment includes:

[0016] Based on the encoder in the multi-material interface control model, the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format are processed to obtain encoding features;

[0017] Based on a fully connected neural network in the multi-substance interface control model, the encoding features are processed to obtain fully connected features;

[0018] Based on the partial derivative calculation network in the multi-material interface control model, performing partial derivative calculation on the fully connected feature to obtain partial derivatives;

[0019] The partial derivatives, the Runge-Kutta coefficients, the target time advancement step and the coding features are processed by a physical information neural network method to determine the volume fraction distribution of the multi-substances.

[0020] In some embodiments of the present disclosure, the processing of the grid unit states on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances to obtain multi-layer grid unit state data on both sides of the multi-substance interface includes:

[0021] Solving the normal direction of the volume fraction distribution of the multi-substances to obtain the normal direction of the interface between the multi-substances;

[0022] constructing a local normal Riemann problem on both sides of the interface between multiple materials based on the normal direction and the time-marching solution result in the second precision format;

[0023] Solving the local normal Riemann problem on both sides of the interface between the multi-materials based on the Riemann solver to obtain state parameters;

[0024] The state parameters are extended along the normal direction to obtain state data of multi-layer grid units on both sides of the multi-material interface.

[0025] In some embodiments of the present disclosure, it also includes:

[0026] Establishing an Euler coordinate system for the multiple substances;

[0027] In the Euler coordinate system, a two-dimensional axis object model is used to determine the initialization parameters of the multi-material.

[0028] In some embodiments of the present disclosure, it also includes:

[0029] The CFL condition is stabilized by using the hyperbolic conservation law, the initialization parameters of the multi-substance are processed, and the target time advancement step is determined.

[0030] In a second aspect, the present disclosure provides a device for determining an interface between multiple substances, the device comprising:

[0031] A first determination module is used to determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and to determine a time advancement solution result in a second precision format based on the initialization parameters of the multi-substances and a target time advancement step size;

[0032] A second determination module is used to input the spatial discretization solution result of the first precision format and the time advancement solution result of the second precision format into the interface control model of multiple substances, solve the interface control model of multiple substances by using a physical information neural network method, and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment;

[0033] An acquisition module is used to process the grid unit states on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances, and acquire multi-layer grid unit state data on both sides of the multi-substance interface;

[0034] The third determination module is used to use the multi-layer grid unit state data on both sides of the multi-substance interface as the interface data between the multi-substances if the current moment does not exceed the preset end time.

[0035] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the device comprising:

[0036] one or more processors;

[0037] a storage device for storing one or more programs,

[0038] When one or more programs are executed by one or more processors, the one or more processors implement the method provided in the first aspect.

[0039] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method provided in the first aspect is implemented.

[0040] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0041] The present invention discloses a method, device, equipment and storage medium for determining the interface between multiple substances. Based on the initialization parameters of multiple substances, a spatial discrete solution result in a first precision format is determined, and based on the initialization parameters of multiple substances and the target time advancement step, a time advancement solution result in a second precision format is determined; the spatial discrete solution result in the first precision format and the time advancement solution result in the second precision format are input into the interface control model of multiple substances, and the interface control model of multiple substances is solved by a physical information neural network method to determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment; the volume fraction distribution of multiple substances is processed by the state of the grid units on both sides of the interface of the multiple substances to obtain the state data of the multi-layer grid units on both sides of the interface of the multiple substances; if the current moment is greater than the preset end time, the state data of the multi-layer grid units on both sides of the interface of the multiple substances is used as the interface data between the multiple substances. Therefore, the physical information neural network (PINNs) method is introduced into the simulation solution of multiple substance interaction, which is used to efficiently and accurately solve the interface control model (VOF model) of multiple substances, realize the accurate and efficient simulation of the problem of multiple substance interaction, and effectively solve the engineering and technical problems related to the field of multi-substance simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0044] Figure 1 A schematic flow chart of a method for determining an interface between multiple substances provided in an embodiment of the present disclosure;

[0045] Figure 2 A schematic diagram of multi-substance interaction provided for an embodiment of the present disclosure;

[0046] Figure 3 A schematic diagram of a two-dimensional axisymmetric model provided in an embodiment of the present disclosure;

[0047] Figure 4 A flow chart of "using a physical information neural network method to solve the interface control model of multiple substances and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment" in S120 provided in an embodiment of the present disclosure;

[0048] Figure 5 A schematic diagram of the structure of a multi-substance interface control model provided by an embodiment of the present disclosure;

[0049] Figure 6 A schematic diagram of the structure of a device for determining interfaces between multiple substances provided in an embodiment of the present disclosure;

[0050] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0051] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0053] Generally speaking, the VOF methods in the existing technology include geometric and numerical methods. The superpixel segmentation (SLIC) algorithm and the piecewise linear (PLIC) algorithm are more common geometric VOF methods, which involve more complex geometric operations, complex logic, large amount of calculation, and are not easy to handle multi-material problems of three or more. The numerical VOF method uses traditional high-precision numerical formats to directly solve the VOF equation, including WENO and THINC formats, which have problems such as large amount of calculation and high numerical dissipation.

[0054] In order to improve the accuracy and efficiency of determining the interface between multiple substances, the following Figures 1 to 5The method for determining the interface between multiple substances provided in the embodiment of the present disclosure is described. In the embodiment of the present disclosure, the method for determining the interface between multiple substances can be executed by an electronic device or a server. A specific explanation is given by taking an electronic device as the execution subject of the method for determining the interface between multiple substances as an example.

[0055] Embodiment 1:

[0056] Figure 1 A schematic flow chart of a method for determining an interface between multiple substances provided in an embodiment of the present disclosure is shown.

[0057] like Figure 1 As shown, the method for determining the interface between multiple substances may include the following steps.

[0058] S110. Determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and determine a time marching solution result in a second precision format based on the initialization parameters of the multi-substances and a target time marching step size.

[0059] It is understandable that in order to distinguish different substances, for example, to distinguish between air and water, it is often necessary to accurately capture the interfaces between multiple substances in order to simulate the interaction problem of multiple substances and thus distinguish different substances.

[0060] In this embodiment, the electronic device first obtains the initialization parameters and target time advancement step of multiple substances, and then solves the mass equation, energy equation, and momentum equation of the Euler equations according to the high-precision solution format. Specifically, it solves the initialization parameters and the target time advancement step to determine the spatial discrete solution result in the first precision format and the time advancement solution result in the second precision format.

[0061] For easier understanding, see Figure 2 Schematic diagram of multi-matter interaction shown in Figure 2 It includes three substances, substance A, substance B, and substance C. Among them, substance A is represented by a black solid sphere, substance B is represented by a hollow sphere, substance C is represented by a gray solid sphere, and the interfaces between substance A and substance B, substance B and substance C, and substance A and substance C are represented by curved dotted lines.

[0062] In this embodiment, the method for determining the initialization parameters of multiple substances includes but is not limited to the following methods: establishing an Euler coordinate system for the multiple substances; and determining the initialization parameters of the multiple substances using a two-dimensional axis object model in the Euler coordinate system.

[0063] Specifically, first, according to the actual physical problem corresponding to multiple materials, the Euler coordinate system is established, and the calculation area size and grid size are defined; then, the initial position of the interface between multiple materials is defined, and the initial physical state of different materials is set, including material type, density, pressure, velocity, volume fraction, etc.; then, for the actual physical problem, it is determined whether it is necessary to set a detonation point, and if necessary, the position of the detonation point and the detonation type are determined; then, according to the actual physical problem, a suitable two-dimensional axisymmetric model is selected, such as the NS equations, the Euler equations, and the fluid elastic-plastic equations, and a parallel method is set according to the actual physical problem, and in the Euler coordinate system, the two-dimensional axisymmetric model is used to determine the initialization parameters of multiple materials, and the Euler equations are used to solve the termination time.

[0064] Optionally, actual physical problems include, but are not limited to, air explosion problems, underwater explosion problems, shaped energy jet problems, bubble jet problems, and sea icebreaking problems.

[0065] For example, a 2D axisymmetric model such as Figure 3 As shown, Figure 3 When the multi-substances include helium and air, the initialization parameters of the multi-substances include but are not limited to a domain length of 450 mm, a height of 90 mm, a grid number of 450*90, a helium bubble diameter of 25 mm, a pressure of 101325 Pa, a density of 0.2228 kg / m3, air 1 is normal pressure air, with a length of 230 mm, a pressure of 101325 Pa, a density of 1.225 kg / m3, air 2 is normal pressure air, with a length of 170 mm, a pressure of 159059 Pa, a density of 1.6861 kg / m3, a speed of -113.534 m / s, and a termination time of 1 ms.

[0066] In this embodiment, the method for determining the target time advancement step size includes, but is not limited to, the following methods: utilizing the hyperbolic conservation law to stabilize the CFL condition, processing the initialization parameters of multiple materials, and determining the target time advancement step size.

[0067] Among them, the target time advancement step refers to the time interval between adjacent moments, and the target time advancement step can be expressed by.

[0068] In this embodiment, the method for determining the spatial discrete solution result in the first precision format includes but is not limited to the following methods: using the third-order WENO format to perform spatial discrete solution on the initialization parameters of multiple materials, and obtaining the solution result in the third-order WENO format as the spatial discrete solution result in the first precision format.

[0069] In this embodiment, the method for determining the time marching solution result in the second precision format includes but is not limited to the following methods: using the second-order RK format, performing time marching solution on the initialization parameters according to the target time marching step, and obtaining the solution result in the second-order RK format as the time marching solution result in the second precision format.

[0070] The time advancement solution results include but are not limited to the density, velocity, pressure and internal energy of different substances after the target time advancement step.

[0071] Therefore, since the spatial discretization solution result is a solution result in the third-order WENO format and the time advancement solution result is a solution result in the second-order RK format, these two solution results belong to high-precision numerical format, which can realize accurate and efficient simulation of multi-material interaction problems.

[0072] S120. Input the spatial discretization solution results in the first precision format and the time advancement solution results in the second precision format into the multi-substance interface control model, solve the multi-substance interface control model using the physical information neural network method, and determine the volume fraction distribution of the multi-substance after the target time advancement step at the current moment.

[0073] In this embodiment, the electronic device introduces the physical information neural network method (PINNs method) into the simulation solution of multi-material interactions, so as to solve the interface control model (VOF equation) of multiple materials based on the spatial discrete solution results and the time advancement solution results in a high-precision numerical format, and obtain the volume fraction distribution of multiple materials after the target time advancement step at the current moment.

[0074] The volume fraction distribution of multiple substances refers to the ratio of the volume of different material objects to the total volume.

[0075] S130, performing grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances, and obtaining multi-layer grid unit state data on both sides of the multi-substance interface.

[0076] In this embodiment, the electronic device uses the real virtual fluid method (RGFM) or the pressure relaxation method to process the grid unit states on both sides of the multi-substance interface for the volume fraction distribution of the multi-substance to determine the multi-layer grid unit state data on both sides of the multi-substance interface.

[0077] In some embodiments, a method for determining state data of multi-layer grid units on both sides of a multi-substance interface using a real virtual fluid method (RGFM) includes, but is not limited to, the following methods: solving the volume fraction distribution of multiple substances in the normal direction to obtain the normal direction of the interface between multiple substances; constructing a local normal Riemann problem on both sides of the interface between multiple substances based on the normal direction and the time-marching solution result in the second precision format; solving the local normal Riemann problem on both sides of the interface between multiple substances based on a Riemann solver to obtain state parameters; and extending the state parameters along the normal direction to obtain state data of multi-layer grid units on both sides of the multi-substance interface.

[0078] In this way, RGFM is used to achieve the effect of accurately determining the state data of multi-layer grid units on both sides of the multi-material interface, which is beneficial to improve the accuracy of determining the interface data between multiple materials.

[0079] Furthermore, after determining the multi-layer grid unit state data on both sides of the multi-substance interface, the electronic device compares the current time with the preset end time. If the current time is less than the preset end time, the step of determining the time advancement solution result in the second precision format to obtain the multi-layer grid unit state data on both sides of the multi-substance interface is re-executed until the recalculated time is greater than the preset end time. The multi-layer grid unit state data on both sides of the multi-substance interface is used as the interface data between the multiple substances to end the numerical calculation. If the current time is greater than the preset end time, S140 is executed to end the numerical calculation.

[0080] S140: If the current time is greater than the preset end time, the state data of the multi-layer grid units on both sides of the multi-substance interface are used as the interface data between the multi-substances.

[0081] In this embodiment, if the electronic device determines that the current time is greater than the preset end time, the numerical calculation is terminated, and the state data of the multi-layer grid units on both sides of the multi-material interface are output as the interface data between the multi-materials, thereby outputting the physical state of the entire area at the current time.

[0082] A method for determining the interface between multiple substances in an embodiment of the present disclosure determines the spatial discrete solution result in a first precision format based on the initialization parameters of the multiple substances, and determines the time advancement solution result in a second precision format based on the initialization parameters of the multiple substances and the target time advancement step; the spatial discrete solution result in the first precision format and the time advancement solution result in the second precision format are input into the interface control model of the multiple substances, and the interface control model of the multiple substances is solved by the physical information neural network method to determine the volume fraction distribution of the multiple substances after the target time advancement step at the current moment; the volume fraction distribution of the multiple substances is processed by the grid unit state on both sides of the multiple substance interface to obtain the multi-layer grid unit state data on both sides of the multiple substance interface; if the current moment is greater than the preset end time, the multi-layer grid unit state data on both sides of the multiple substance interface is used as the interface data between the multiple substances. Therefore, the physical information neural network (PINNs) method is introduced into the simulation solution of multiple substance interaction, which is used to efficiently and accurately solve the interface control model (VOF model) of multiple substances, realize the accurate and efficient simulation of the problem of multiple substance interaction, and effectively solve the engineering and technical problems related to the field of multi-substance simulation.

[0083] In another embodiment of the present disclosure, the process of "using the physical information neural network method to solve the interface control model of multiple substances and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment" in S120 is explained in detail.

[0084] Figure 4 A flow chart of "using a physical information neural network method to solve the interface control model of multiple substances and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment" in S120 provided by an embodiment of the present disclosure is shown.

[0085] like Figure 4 As shown, in S120, "using the physical information neural network method to solve the interface control model of multiple substances and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment" may include the following steps.

[0086] S410. Based on the encoder in the multi-material interface control model, the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format are processed to obtain coding features.

[0087] In this embodiment, the multi-substance interface control model can be a neural network model trained based on the VOF model. Specifically, the multi-substance interface control model includes an encoder, a fully connected neural network, a partial derivative calculation network, and a module integrating a physical information neural network method.

[0088] For easier understanding, see Figure 5The schematic diagram of the structure of the multi-material interface control model is shown. The electronic device will obtain the spatial discretization solution result in the first precision format. and time-marching solution results in second precision format Input the multi-material interface control model, based on which the encoder and Perform encoding processing to obtain encoding features .

[0089] S420. Based on a fully connected neural network in a multi-material interface control model, the coding features are processed to obtain fully connected features.

[0090] In this embodiment, the electronic device randomly selects values ​​within a certain range, including the target time advancement step. ,speed , coordinate point , and the step length ,speed , coordinate point And the encoding features Input the fully connected neural network in the multi-material interface control model, which adopts The Runge-Kutta method of level 1 is used to process the input features, and the output length is Fully connected features .

[0091] S430. Based on the partial derivative calculation network in the multi-material interface control model, partial derivatives are calculated for the fully connected features to obtain partial derivatives.

[0092] In this embodiment, the electronic device passes through the partial derivative calculation network in the multi-material interface control model, and uses the automatic differentiation (AD) technology of the neural network to obtain the partial derivative through back propagation. .

[0093] S440. Physical information neural network method is used to process partial derivatives, Runge-Kutta coefficients, target time advancement steps and coding features to determine the volume fraction distribution of multiple materials.

[0094] In this embodiment, the electronic device obtains the partial derivative , Runge-Kutta coefficient (parameters of the Runge-Kutta method) , target time advancement step , velocity field u, and then use the physical information neural network method (PINNs method) to process and determine the volume fraction distribution of multiple materials .

[0095] Based on different network layers in the multi-material interface control model and combined with the physical information neural network method, the complex interface between multiple materials can be accurately captured, which can suppress non-physical dissipation and oscillation and has good conservation characteristics, thus effectively meeting the needs of accurate simulation of interactions between multiple materials.

[0096] In this embodiment, the electronic device can use the minimized mean square error loss function (MSE) to iteratively train the initial neural network, and obtain a trained multi-material interface control model when the value of the minimized mean square error loss function (MSE) meets a preset condition.

[0097] Alternatively, the volume fraction distribution of multiple materials can be determined as follows:

[0098] ;

[0099] in, , and are the volume fractions at the current moment, is the volume fraction required at the next moment, is the target time advancement step, , , is the Runge-Kutta coefficient (parameter of the Runge-Kutta method), is the characteristic of the neural network, specifically the volume fraction between the current moment and the next moment, which is determined by the time coefficient of RK. Substitute it into the calculation to obtain the volume fraction of multiple substances at the current moment.

[0100] Minimizing the mean square error loss function (MSE) can be expressed as follows:

[0101] ;

[0102] in, is the volume fraction at the current moment, Substitute the formula to calculate the volume fraction of multiple substances at the current moment, is the number of coordinate points, are the coordinates of the training data.

[0103] Embodiment 2:

[0104] The embodiment of the present disclosure also provides a device for determining a multi-substance interface for implementing the above-mentioned method for determining a multi-substance interface. Figure 6 In the embodiment of the present disclosure, the device for determining the interface between multiple substances may be an electronic device or a server.

[0105] Figure 6 A schematic structural diagram of a device for determining interfaces between multiple substances provided in an embodiment of the present disclosure is shown.

[0106] like Figure 6 As shown, the device 600 for determining the interface between multiple substances may include:

[0107] A first determination module 610 is used to determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and to determine a time marching solution result in a second precision format based on the initialization parameters of the multi-substances and a target time marching step size;

[0108] A second determination module 620 is used to input the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of multiple materials, solve the interface control model of multiple materials by using a physical information neural network method, and determine the volume fraction distribution of multiple materials after the target time advancement step at the current moment;

[0109] An acquisition module 630 is used to perform grid unit state processing on both sides of the multi-substance interface on the volume fraction distribution of the multi-substances, and acquire multi-layer grid unit state data on both sides of the multi-substance interface;

[0110] The third determination module 640 is configured to use the multi-layer grid unit state data on both sides of the multi-substance interface as the interface data between the multi-substances if the current time is greater than a preset end time.

[0111] The device for determining the interface between multiple substances in the embodiment of the present disclosure determines the spatial discrete solution result in the first precision format based on the initialization parameters of the multiple substances, and determines the time advancement solution result in the second precision format based on the initialization parameters of the multiple substances and the target time advancement step; inputs the spatial discrete solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of the multiple substances, solves the interface control model of the multiple substances by using the physical information neural network method, and determines the volume fraction distribution of the multiple substances after the target time advancement step at the current moment; processes the state of the grid units on both sides of the interface of the multiple substances to obtain the state data of the multi-layer grid units on both sides of the interface of the multiple substances; if the current moment is greater than the preset end time, the state data of the multi-layer grid units on both sides of the interface of the multiple substances is used as the interface data between the multiple substances. Therefore, the physical information neural network (PINNs) method is introduced into the simulation solution of the interaction between multiple substances, which is used to efficiently and accurately solve the interface control model (VOF model) of multiple substances, realize the accurate and efficient simulation of the interaction problem of multiple substances, and effectively solve the engineering and technical problems related to the field of multi-substance simulation.

[0112] In some embodiments of the present disclosure, the first determining module 610 includes:

[0113] The spatial discrete solution unit is used to perform spatial discrete solution on the initialization parameters of the multi-materials in a third-order WENO format, and obtain the solution result in the third-order WENO format as the spatial discrete solution result in the first precision format.

[0114] In some embodiments of the present disclosure, the first determining module 610 includes:

[0115] The time-marching solution unit is used to use the second-order RK format to perform time-marching solution on the initialization parameters according to the target time-marching step size, and obtain the solution result in the second-order RK format as the time-marching solution result in the second precision format.

[0116] In some embodiments of the present disclosure, the second determination module 620 includes:

[0117] An encoding unit, configured to process the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format based on an encoder in the multi-substance interface control model to obtain encoding features;

[0118] A fully connected processing unit, used to process the encoding features based on the fully connected neural network in the multi-substance interface control model to obtain fully connected features;

[0119] A partial derivative calculation unit, configured to calculate partial derivatives of the fully connected features based on a partial derivative calculation network in the multi-substance interface control model to obtain partial derivatives;

[0120] The volume fraction determination unit is used to process the partial derivatives, the Runge-Kutta coefficients, the target time advancement step and the coding features using a physical information neural network method to determine the volume fraction distribution of the multiple substances.

[0121] In some embodiments of the present disclosure, the acquisition module 630 includes:

[0122] A normal solving unit, used for solving the normal direction of the volume fraction distribution of the multi-substances to obtain the normal direction of the interface between the multi-substances;

[0123] A construction unit, configured to construct a local normal Riemann problem on both sides of the interface between multiple materials based on the normal direction and the time-marching solution result in the second precision format;

[0124] A solving unit, used for solving the local normal Riemann problem on both sides of the interface between the multi-materials based on a Riemann solver to obtain state parameters;

[0125] The acquisition unit is used to extend the state parameter along the normal direction to obtain the state data of the multi-layer grid units on both sides of the multi-material interface.

[0126] In some embodiments of the present disclosure, the device further includes:

[0127] A coordinate system establishment module, used for establishing an Euler coordinate system for the multi-substances;

[0128] The initialization parameter determination module is used to determine the initialization parameters of the multi-substances by using a two-dimensional axis object model in the Euler coordinate system.

[0129] In some embodiments of the present disclosure, the device further includes:

[0130] The time advancement step determination module is used to utilize the hyperbolic conservation law to stabilize the CFL condition, process the initialization parameters of the multi-substances, and determine the target time advancement step.

[0131] It should be noted that Figure 6 The apparatus 600 for determining the interface between multiple substances can be executed Figures 1 to 5 The various steps in the method embodiment shown in the figure are implemented Figures 1 to 5 The various processes and effects in the method embodiment shown are not described in detail here.

[0132] Embodiment three:

[0133] Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown.

[0134] like Figure 7 As shown, the electronic device may include a processor 701 and a memory 702 storing computer program instructions.

[0135] Specifically, the processor 701 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0136] The memory 702 may include a large capacity memory for information or instructions. By way of example and not limitation, the memory 702 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a tape, or a universal serial bus (USB) drive or a combination of two or more of these. Where appropriate, the memory 702 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 702 may be inside or outside the integrated gateway device. In a specific embodiment, the memory 702 is a non-volatile solid-state memory. In a specific embodiment, the memory 702 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (Electrically Erasable Programmable ROM, EEPROM), an electrically rewritable ROM (EAROM) or a flash memory, or a combination of two or more of these.

[0137] The processor 701 reads and executes the computer program instructions stored in the memory 702 to perform the steps of the method for determining the interface between multiple substances provided in the embodiment of the present disclosure.

[0138] In one example, the electronic device may further include a transceiver 703 and a bus 704. Figure 7 As shown, the processor 701, the memory 702 and the transceiver 703 are connected via a bus 704 and communicate with each other.

[0139] The bus 704 includes hardware, software, or both. For example, but not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a Memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, the bus 704 may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0140] The following is an embodiment of a computer-readable storage medium provided in an embodiment of the present disclosure. The computer-readable storage medium and the method for determining the interface between multiple substances in the above-mentioned embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the computer-readable storage medium, reference can be made to the embodiment of the method for determining the interface between multiple substances in the above-mentioned embodiments.

[0141] This embodiment provides a storage medium containing computer executable instructions. When the computer executable instructions are executed by a computer processor, they are used to perform a method for determining an interface between multiple substances. The method includes:

[0142] Determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and determine a time marching solution result in a second precision format based on the initialization parameters of the multi-substances and a target time marching step size;

[0143] Inputting the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of multiple materials, solving the interface control model of multiple materials by using the physical information neural network method, and determining the volume fraction distribution of multiple materials after the target time advancement step at the current moment;

[0144] Performing grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances to obtain multi-layer grid unit state data on both sides of the multi-substance interface;

[0145] If the current time is greater than the preset end time, the state data of the multi-layer grid units on both sides of the multi-substance interface are used as the interface data between the multi-substances.

[0146] Of course, the computer executable instructions of a storage medium including computer executable instructions provided in an embodiment of the present disclosure are not limited to the above method operations, and can also execute related operations in the method for determining the interface between multiple substances provided in any embodiment of the present disclosure.

[0147] Through the above description of the implementation methods, the technical personnel in the relevant field can clearly understand that the present disclosure can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present disclosure, or the part that contributes to the prior art, can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FLASH), a hard disk or an optical disk, etc., including a number of instructions for enabling a computer cloud platform (which can be a personal computer, a server, or a network cloud platform, etc.) to execute the method for determining the interface between multiple substances provided in each embodiment of the present disclosure.

[0148] Note that the above are only preferred embodiments of the present disclosure and the technical principles used. Those skilled in the art will understand that the present disclosure is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present disclosure. Therefore, although the present disclosure is described in more detail through the above embodiments, the present disclosure is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present disclosure, and the scope of the present disclosure is determined by the scope of the appended claims.

Claims

1. A method for determining an interface between multiple substances, characterized in that: include: Determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and determine a time marching solution result in a second precision format based on the initialization parameters of the multi-substances and a target time marching step size; Inputting the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the interface control model of multiple materials, solving the interface control model of multiple materials by using the physical information neural network method, and determining the volume fraction distribution of multiple materials after the target time advancement step at the current moment; Performing grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances to obtain multi-layer grid unit state data on both sides of the multi-substance interface; If the current time is greater than the preset end time, the state data of the multi-layer grid units on both sides of the multi-substance interface are used as the interface data between the multi-substances.

2. The method according to claim 1, characterized in that The step of determining the spatial discretization solution result in the first precision format based on the initialization parameters of the multi-material comprises: A third-order WENO format is used to perform spatial discretization solution on the initialization parameters of the multi-material, and a solution result in the third-order WENO format is obtained as a spatial discretization solution result in the first precision format.

3. The method according to claim 1, characterized in that: The step of determining a time-marching solution result in a second precision format based on the initialization parameters of the multi-material and the target time-marching step size includes: The second-order RK format is adopted to perform a time-marching solution on the initialization parameters according to the target time-marching step, and a solution result in the second-order RK format is obtained as the time-marching solution result in the second precision format.

4. The method according to claim 1, characterized in that: The step of inputting the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format into the multi-substance interface control model, solving the multi-substance interface control model by using a physical information neural network method, and determining the volume fraction distribution of the multi-substance after the target time advancement step at the current moment includes: Based on the encoder in the multi-material interface control model, the spatial discretization solution result in the first precision format and the time advancement solution result in the second precision format are processed to obtain encoding features; Based on a fully connected neural network in the multi-substance interface control model, the encoding features are processed to obtain fully connected features; Based on the partial derivative calculation network in the multi-material interface control model, performing partial derivative calculation on the fully connected feature to obtain partial derivatives; The partial derivatives, the Runge-Kutta coefficients, the target time advancement step and the coding features are processed by a physical information neural network method to determine the volume fraction distribution of the multi-substances.

5. The method according to claim 1, characterized in that The step of performing grid unit state processing on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances to obtain multi-layer grid unit state data on both sides of the multi-substance interface includes: Solving the normal direction of the volume fraction distribution of the multi-substances to obtain the normal direction of the interface between the multi-substances; constructing a local normal Riemann problem on both sides of the interface between multiple materials based on the normal direction and the time-marching solution result in the second precision format; Solving the local normal Riemann problem on both sides of the interface between the multi-materials based on the Riemann solver to obtain state parameters; The state parameters are extended along the normal direction to obtain state data of multi-layer grid units on both sides of the multi-material interface.

6. The method according to claim 1, characterized in that Also includes: Establishing an Euler coordinate system for the multiple substances; In the Euler coordinate system, a two-dimensional axis object model is used to determine the initialization parameters of the multi-material.

7. The method according to claim 1, characterized in that Also includes: The CFL condition is stabilized by using the hyperbolic conservation law, the initialization parameters of the multi-substance are processed, and the target time advancement step is determined.

8. A device for determining interfaces between multiple substances, characterized in that: include: A first determination module is used to determine a spatial discretization solution result in a first precision format based on the initialization parameters of the multi-substances, and to determine a time advancement solution result in a second precision format based on the initialization parameters of the multi-substances and a target time advancement step size; A second determination module is used to input the spatial discretization solution result of the first precision format and the time advancement solution result of the second precision format into the interface control model of multiple substances, solve the interface control model of multiple substances by using a physical information neural network method, and determine the volume fraction distribution of multiple substances after the target time advancement step at the current moment; An acquisition module is used to process the grid unit states on both sides of the multi-substance interface for the volume fraction distribution of the multi-substances, and acquire multi-layer grid unit state data on both sides of the multi-substance interface; The third determination module is used to use the multi-layer grid unit state data on both sides of the multi-substance interface as the interface data between the multi-substances if the current time is greater than a preset end time.

9. An electronic device, characterized in that: include: processor; A memory for storing executable instructions; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the method according to any one of claims 1 to 7.