Method and device for acquiring physical property parameters during mesh deformation

Abstract

The invention discloses a method and a device for acquiring physical property parameters during grid deformation, and relates to the technical field of phase change problem numerical simulation. The method comprises the following steps: establishing a geometric model according to size parameters of a multi-physical-property-parameter material; grid division is carried out on the geometric model, an initial grid is determined, and corresponding material numbers are distributed to units of the initial grid according to the physical parameters of the multi-physical-parameter material; processing the initial grid by adopting a grid deformation method to obtain a deformed grid; and a parameter mapping model is determined based on fitting of the node coordinates of the units of the initial grid and the material numbers, and the corresponding material numbers are determined based on the node coordinates of the deformed grid through the parameter mapping model. Based on the mapping mode, the physical property parameters are efficiently and accurately obtained after the grid is deformed, and the reliability of obtaining the physical property parameters under the grid deformation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of numerical simulation of phase change problems, and particularly to a method and device for obtaining physical property parameters during grid deformation. Background Art

[0002] In many engineering fields, phase change problems such as ablation and aircraft icing play crucial roles. Exemplarily, in the field of manned spaceflight, when reentry warheads, manned spacecraft return capsules and other reentry vehicles pass through the atmosphere at hypersonic speeds, the surface temperature will rise sharply due to aerodynamic heating. To prevent heat from being transferred to the interior of the vehicle and damaging the astronauts and equipment, ablation heat protection is considered. By melting or vaporizing the ablation material covering the surface of the vehicle, that is, using the phase change process of the material to absorb a large amount of aerodynamic heat and prevent aerodynamic heat from being conducted to the interior of the vehicle, as Figure 1 shown. Figure 1 (a) shows the state of a certain ablation heat protection system before ablation. Figure 1 (b) shows the situation after the ablation process.

[0003] When numerically simulating problems such as heat transfer and deformation of substances with phase changes, it is often necessary to simulate the heat transfer field, deformation field or other related physical fields inside the solid. To handle these complex physical phenomena, grid discretization technology is generally used. The grid serves as the basis for discretization, and the grid nodes serve as the storage locations of the discretized physical quantities. The grid is used to discretize the spatial positions occupied by the solid. The initial grid covers the geometric shape and spatial occupancy area of the object before ablation or icing.

[0004] Current methods for these problems with moving boundaries include grid deformation methods and non-grid deformation methods. When using non-grid deformation methods to handle phase change problems, there is no need to consider the acquisition of physical property parameters after phase change. However, the interface capture is inaccurate, and the formed interface has non-physical corners, pits and unevenness, which affects the application of heat flux boundary conditions and reduces the accuracy of simulation results. Moreover, the grids generated by this method cannot conform to the movement of the object surface, and the grids in the normal direction are not dense after phase change, which will also affect the calculation accuracy. Compared with non-grid deformation methods, grid deformation methods use grid deformation methods to adjust the grids used in numerical simulation, so that the area occupied by the grids changes with the change of the material shape. When dealing with phase change problems, the interface capture is more accurate, and the generated body-fitted grids can move along with the object surface. After phase change occurs, the grid density in the normal direction will not change and will not affect the accuracy of the calculation results. However, during the phase change process, the corresponding relationship between the grid elements and the substance changes due to grid deformation. To improve the reliability of phase change process simulation, it is necessary to ensure the precise matching of physical property parameters. Therefore, how to improve the reliability of obtaining physical property parameters caused by grid deformation becomes a problem to be solved. Summary of the Invention

[0005] The present invention provides a method and device for obtaining physical property parameters during grid deformation, and solves the technical problem of improving the reliability of obtaining physical property parameters caused by grid deformation.

[0006] A method for obtaining physical property parameters during grid deformation provided by the first aspect of the present invention includes: Establish a geometric model according to the dimensional parameters of the multi-physical property parameter material; Perform grid division on the geometric model to determine the initial grid, and assign corresponding material numbers to the cells of the initial grid according to the physical property parameters of the multi-physical property parameter material; Process the initial grid by using a grid deformation method to obtain a deformed grid; Based on the node coordinates and material numbers of the cells of the initial grid, fit to determine a parameter mapping model, and determine the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model.

[0007] Further, the process of determining the parameter mapping model based on the fitting of the node coordinates and material numbers of the cells of the initial grid, and determining the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model includes: Use the node coordinates and corresponding material numbers of each cell of the initial grid, and fit them by using an interpolation function, and construct the obtained interpolation system as the parameter mapping model; Input the node coordinates of the deformed grid into the parameter mapping model, and output the corresponding interpolation results; Use the material number closest to each interpolation result as the material number corresponding to each cell of the deformed grid.

[0008] Further, the process of determining the parameter mapping model based on the fitting of the node coordinates and material numbers of the cells of the initial grid, and determining the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model includes: Input the node coordinates of any cell of the initial grid into the initial artificial neural network model, and output the predicted material number; Calculate the loss function value by using the predicted material number and the corresponding material number, and iteratively optimize the initial artificial neural network model based on the loss function value to determine the parameter mapping model; Input the node coordinates of each cell of the deformed grid into the parameter mapping model, and output the material numbers corresponding to each cell of the deformed grid.

[0009] Further, the calculation process of the loss function value includes:

[0010] Wherein, is the loss function, is the total number of nodes in the grid, is the node index of the grid , is the th node coordinate, is the variable model parameter, is the predicted material number, is the th material number corresponding to the node coordinate.

[0011] Furthermore, the interpolation function is a radial basis interpolation function.

[0012] Furthermore, the initial artificial neural network model is a fully connected neural network model.

[0013] A device for obtaining physical property parameters during grid deformation provided by the second aspect of the present invention includes: A geometric modeling module, configured to establish a geometric model according to the size parameters of the multi-physical property parameter material; A grid division module, configured to divide the geometric model into grids to determine an initial grid, and assign corresponding material numbers to the cells of the initial grid according to the physical property parameters of the multi-physical property parameter material; A grid deformation module, configured to process the initial grid by using a grid deformation method to obtain a deformed grid; A deformation parameter acquisition module, configured to fit and determine a parameter mapping model based on the node coordinates and material numbers of the cells belonging to the initial grid, and determine the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model.

[0014] A computer device provided by the third aspect of the present invention includes a memory and a processor. When the computer program stored in the memory is executed by the processor, the processor is caused to execute the steps of the method for obtaining physical property parameters during grid deformation as described in any one of the above.

[0015] A computer-readable storage medium provided by the fourth aspect of the present invention stores a computer program thereon. When the computer program is executed, the method for obtaining physical property parameters during grid deformation as described in any one of the above is implemented.

[0016] A computer program product provided by the fifth aspect of the present invention includes a computer program / instructions. When the computer program / instructions are executed by a processor, the method for obtaining physical property parameters during grid deformation as described in any one of the above is implemented.

[0017] It can be seen from the above technical solutions that the present invention has the following advantages: The above solution of the present invention provides a method for obtaining physical property parameters during grid deformation, including: establishing a geometric model according to the size parameters of a multi-physical property parameter material; performing grid division on the geometric model to determine the initial grid, and assigning corresponding material numbers to the units of the initial grid according to the physical property parameters of the multi-physical property parameter material; processing the initial grid by using a grid deformation method to obtain the deformed grid; fitting to determine a parameter mapping model based on the node coordinates and material numbers to which the units of the initial grid belong, and determining the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model. A corresponding relationship between the grid node coordinates and the physical property parameters is established based on the mapping method, so as to obtain the parameter mapping model, and the physical property parameters of the deformed grid cells are obtained based on the parameter mapping model after the grid deformation, without repeating the cumbersome data processing work every time after deformation, and the physical property parameters can be obtained efficiently and accurately in complex phase change numerical simulations, thereby improving the reliability of obtaining physical property parameters under grid deformation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a schematic diagram of the comparison before and after ablation of a certain thermal protection system of an existing aircraft; Figure 2 It is a flowchart of the steps of a method for obtaining physical property parameters during grid deformation provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the comparison of the regions and grids of a certain multi-physical property parameter material before and after ablation provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of moving boundary processing provided by an embodiment of the present invention; Figure 5 It is a schematic diagram of the principle of the grid reconstruction method provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of the principle of the raster method provided by an embodiment of the present invention; Figure 7 It is a schematic diagram of ablation treatment of quadrilateral elements provided by an embodiment of the present invention; Figure 8 It is a schematic diagram of the comparison between body-fitted grids and Cartesian grids provided by an embodiment of the present invention; Figure 9 It is a schematic diagram of the comparison before and after icing of an aircraft wing provided by an embodiment of the present invention; Figure 10Schematic diagram of the comparison between the regions before and after icing on the aircraft airfoil provided by the embodiment of the present invention; Figure 11 Schematic diagram of the correspondence between the multi-physical property parameter material coordinates and the physical property parameters provided by the embodiment of the present invention; Figure 12 Schematic diagram of the multi-physical property parameter material of the three-dimensional test case provided by the embodiment of the present invention; Figure 13 Schematic diagram of the code running result of the three-dimensional test case provided by the embodiment of the present invention; Figure 14 Schematic diagram of the calculation process of the fully connected neural network provided by the embodiment of the present invention; Figure 15 Schematic diagram of the calculation process of a single neuron provided by the embodiment of the present invention; Figure 16 Schematic diagram of the process of obtaining physical property parameters by the mapping method for multi-physical property parameter materials under grid deformation conditions provided by the embodiment of the present invention; Figure 17 Structural block diagram of a device for obtaining physical property parameters during grid deformation provided by the embodiment of the present invention. Detailed implementation manners

[0020] The embodiment of the present invention provides a method and a device for obtaining physical property parameters during grid deformation, which are used to solve the technical problem of improving the reliability of obtaining physical property parameters caused by grid deformation.

[0021] To make the objectives, features, and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0022] Term explanations Physical property parameters: Parameters or indicators that describe the properties and characteristics of substances, such as density, thermal conductivity, etc.

[0023] Multi-layer thermal protection system: Composed of a heat protection layer, a heat insulation layer, and a load-bearing structure. Compared with the conventional thermal protection structure, it can adapt to a more complex aerodynamic heat environment and protect the aircraft.

[0024] Grid deformation: During numerical simulations such as computational fluid dynamics (CFD) or finite element analysis, the deformation of the grid during the simulation process as the shape of the object or the physical process changes.

[0025] Phase change: It refers to the process in which a substance changes from one phase (state of matter) to another under certain conditions.

[0026] Ablation: It refers to the phenomenon of surface mass loss of an object caused by thermochemical or mechanical processes under the action of high-temperature gas flow, such as melting, combustion, and sublimation.

[0027] Aircraft icing: Aircraft icing refers to the phenomenon that certain parts of the aircraft surface freeze during flight, especially when flying through supercooled water droplet (liquid water with a temperature below the freezing point) clouds.

[0028] Moving boundary problem: It refers to the phenomenon in fluid flow or other physical processes where the boundary position changes with time.

[0029] Mesh reconstruction: Mesh reconstruction is a method of regenerating a mesh after the structure shape changes.

[0030] Element birth and death method: The effect of mesh elimination is simulated by marking certain elements as "dead" (i.e., not considering their contribution to the structural response). These elements are ignored in the analysis, thus reducing the computational complexity.

[0031] Immersed boundary method: The interaction between the object boundary and the fluid is simulated by adding a body force term to the fluid motion equation. This method allows calculations on a simple Cartesian grid, avoiding the difficulty of generating body-fitted meshes for complex shapes.

[0032] Grid method: A technique for handling mesh reconstruction in ablation simulation.

[0033] Robustness: It refers to the ability of a system, algorithm, or method to maintain stability and good performance in the face of uncertainty, perturbation, error, or change.

[0034] Interface tracking: Used to accurately capture and describe the interface between the fluid and the solid boundary.

[0035] Interpolation system: A mathematical tool in numerical analysis for estimating the values of unknown data points between known data points.

[0036] Radial basis function: A special real-valued function whose value depends only on the distance between the input vector and a certain center point (or the origin).

[0037] Artificial neural network fitting: A function approximation technique based on artificial neural networks. It trains the neural network to learn the mapping relationship between the input data and the output data, and then predicts and interpolates the unknown data points.

[0038] Mapping: A mathematical term referring to a relationship between two sets that associates each element in the first set with an element in the second set.

[0039] Please refer to Figure 2 , Figure 2 which is the flowchart of the steps of a method for obtaining physical property parameters during grid deformation provided by an embodiment of the present invention.

[0040] A method for obtaining physical property parameters during grid deformation provided by this embodiment includes: Step 101: Establish a geometric model according to the size parameters of the multi-physical property parameter material.

[0041] A multi-physical property parameter material refers to a material with different physical property parameters in different parts of the material. Exemplarily, a multi-physical property parameter material can be understood as an object composed of multiple materials. For example, in order to adapt to the increasingly harsh aerodynamic heat environment of aerospace vehicles and achieve the goal of reducing the weight of the vehicle, in recent years, a multi-layer thermal protection system composed of multi-layer thermal protection materials has been proposed for thermal protection, and multi-layer materials mean multiple physical property parameters, such as Figure 3 as shown in (a), different color blocks in the figure represent different materials. The outermost layer is composed of three different materials, namely Material 1, Material 2, and Material 3, and the second layer is composed of another material, namely Material 4, and each material has different physical property parameters.

[0042] Size parameters refer to the parameters that describe the geometric shape characteristics of an object.

[0043] A geometric model refers to a mathematical model that abstractly represents the geometric shape characteristics of an object.

[0044] It should be noted that in this embodiment, a corresponding geometric model is constructed according to the size parameters of the multi-physical property parameter material, which is the geometric basis for numerical simulation.

[0045] Step 102: Perform grid division on the geometric model to determine the initial grid, and assign corresponding material numbers to the cells of the initial grid according to the physical property parameters of the multi-physical property parameter material.

[0046] The initial grid refers to the grid structure composed of multiple cells before phase change after discretizing the research object. A cell is the basic unit of the grid, and cells are connected by a group of nodes.

[0047] The material number refers to a specific label representing a certain physical property parameter, including but not limited to numerical number form and parameter information form, etc.

[0048] It should be noted that, in order to better simulate the phase change situation, before the phase change occurs, the geometric model is discretized into multiple grid cells through grid division, thereby obtaining the initial grid. Different physical property parameters in the multi-physical property parameter material are identified and distinguished by corresponding labels, which are called material numbers in this embodiment. According to the distribution of physical property parameters in the multi-physical property parameter material, the material numbers corresponding to the physical property parameters associated with each unit of the initial grid are assigned to establish the mapping relationship between each unit and its corresponding physical property parameters before the phase change occurs.

[0049] Step 103: Process the initial grid using a grid deformation method to obtain a deformed grid.

[0050] The grid deformation method refers to the grid deformation technology based on grids, such as the grid deformation method based on the finite element method, the grid deformation method based on the interpolation algorithm, etc. For details, reference can be made to the prior art and will not be elaborated here.

[0051] It can be understood that the non-grid deformation methods opposite to the grid deformation method include the birth and death element method, the grid reconstruction method, and the immersed boundary method; The birth and death element method simulates the effect of grid elimination by marking some elements as "dead". Taking the ablation simulation of tetrahedral elements using the birth and death element method as an example, first calculate the ablation thickness of the material at the current moment (i.e., Figure 4 the distance between the upper and lower dotted lines in, the upper dotted line is the initial interface, and the lower dotted line is the interface after ablation), judge the elements that are completely within the ablation thickness (i.e., Figure 4 the white area elements in), and eliminate them. The remaining black area elements are the parts remaining after ablation. Subsequently, use the temperature field at the previous moment as the current initial temperature field, and apply heat flux on the new surface to update the temperature field distribution; Grid reconstruction is a method of regenerating grids after the structure shape changes. Taking ablation simulation as an example, as Figure 5 shown, Figure 5 (a) is the grid covering the area before ablation, Figure 5 (b) the gray area in is the area where the ablated material is located. This method first calculates the shape after ablation, and then regenerates a set of grids according to the new shape, that is, the grid shown in Figure 5 (b); due to the large amount of calculation for the overall grid reconstruction, the local grid reconstruction based on the grid method that only reconstructs locally near the ablation boundary is widely used. As Figure 6 shown, Figure 6 (a) is the grid covering the area before ablation, Figure 6 (b) the gray area in is the area where the ablated material is located. This method first judges the grids outside the ablation area and discards them, and retains the grids that are completely within the ablation area and intersect with its boundary, that is, the grids shown in Figure 6 (b). Then, process the grids intersecting with the ablation boundary. Assume that the critical temperature at a certain moment is T sis 800 °C. The part exceeding this critical temperature is burned off while the part below it is retained. Figure 7 (a) is Figure 6 (b) is an enlarged view of a certain boundary grid. The temperatures of nodes 1, 2, 3, and 4 are 700 °C, 980 °C, 600 °C, and 400 °C respectively. New nodes 5 and 6 corresponding to the critical temperature are interpolated on the unit edge (as shown in Figure 7 (b)). The area enclosed by nodes 1, 5, 6, 3, and 4 represents the area where the temperature is lower than T s . This polygon area needs to be reconstructed, as shown in Figure 7 (c). The polygon area can be split into three triangular elements. When the area to be reconstructed is a quadrilateral, it is split into two triangular elements. If it is a triangular element, no processing is required. Different from other methods where the fluid grid and the solid grid are divided separately, the immersed boundary method uses a set of grids to cover both the fluid and solid regions simultaneously, as shown in Figure 8 . Figure 8 In (b), the dark line is the shape of the target object, and the right-angled grid formed by the light lines is the Cartesian grid. The entire flow field calculation uses the Cartesian grid, rather than the complex body-fitted grid generated according to the object shape as shown in Figure 8 (a). When dealing with the moving boundary problem, the boundary conditions of the fluid on the solid are applied through a certain force field. The fluid and solid regions are distinguished by a certain form (such as a numbering system). When the fluid and solid regions change, such as Figure 8 the gray area in (c) is Figure 8 the ablated outer shape in (b). This form also changes accordingly, but the grid remains unchanged. Figure 8 (b) and Figure 8 (c) have the same grid. This method does not need to deal with the coordinate and grid conversion from the physical plane to the computational plane. It has significant advantages in cases where the boundary is complex or there are multiple immersed bodies, especially in dynamic boundary problems. There is no need to update the grid in real time at each time step, which greatly improves the calculation efficiency and saves the grid generation time.

[0052] The deformed grid refers to the grid structure after the cells of the initial grid have changed after a phase change.

[0053] It should be noted that in this embodiment, the grid deformation method is used to simulate the phase change of the initial grid, thereby obtaining the corresponding deformed grid. It can be understood that when a material undergoes a phase change, such as the surface recession caused by ablation or the surface uplift caused by icing, or other phase change problems that cause changes in the solid region, using the grid deformation technology to handle it is to move the initial grid cells of the initial grid so that the area it covers deforms into the area after ablation or ice uplift. However, due to the occurrence of phase change, the corresponding relationship between the grid and the physical properties parameters has changed; exemplarily, as Figure 3 shown, due to ablation of the aircraft in a severe aerodynamic heating environment resulting in material recession, Figure 3 the marked grid cells (the black grid cells in the figure) of the initial grid before ablation shown in (a) are located in material 2 region. According to Figure 3 the deformed grid after ablation shown in (b), it can be seen that after the grid deformation, the marked grid cells have run to material 4 region; exemplarily, when an aircraft flies through a cloud region, cloud water droplets impact the aircraft surface to form a water film. In a low-temperature environment, these water films will undergo a phase change to transform into ice and gradually accumulate, forming ice layers or ice crystals especially at the leading edge of the wing on the aircraft surface. The weight of the wing increases, and the lift increases, which increases the surface roughness of the wing and thus changes the streamline shape of the wing, as Figure 9 shown, Figure 9 (a) is a schematic diagram before the wing is iced, Figure 9 (b) is a schematic diagram after the wing is iced. And this will cause the air flow velocity flowing through the wing surface to decrease and separate in advance, thereby reducing the lift and increasing the drag coefficient, and reducing the maximum lift-drag ratio. When the lift is not sufficient to support the weight of the aircraft, the danger of the aircraft will increase; when using the grid deformation technology to deal with the icing problem, there will also be a problem that the corresponding relationship between the grid cells and the substance is known before the phase change occurs, and the parameter correspondence is disordered after the grid deformation, as Figure 10 shown, Figure 10 the marked grid cells (the black grid cells in the figure) before icing in (a) are located in the aircraft structure region, and according to Figure 10 (b), it can be seen that it is located in the ice layer region after icing.

[0054] Step 104: Determine the parameter mapping model by fitting the node coordinates and material numbers to which the cells of the initial grid belong, and determine the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model.

[0055] Node coordinates refer to the position parameters describing the cells of the grid in a specific coordinate system, including but not limited to two-dimensional or three-dimensional forms.

[0056] The parameter mapping model refers to a mapping model with the node coordinates of the grid as input parameters and the material numbers as output parameters, in which a mapping relationship between the node coordinates of the grid and the material numbers is established, and it can realize the inference or prediction from known conditions to unknown results.

[0057] It should be noted that in this embodiment, the node coordinates to which the cells of each initial grid belong and their corresponding material numbers are used as known conditions, and the mapping relationship between the node coordinates of the grid and the material numbers is established through parameter fitting, so as to determine the parameter mapping model; this mapping relationship can be referred to Figure 11 shown, where represents the material number corresponding to the grid node coordinates. Taking two dimensions as an example, the node coordinates of the grid , is the X-axis coordinate value, is the Y-axis coordinate value; To track and match the physical property parameters of each element in the deformed grid, first extract the node coordinates of each element in the deformed grid and input them into the parameter mapping model in sequence. In the parameter mapping model, based on the established mapping relationship, output the material numbers of the node coordinates. According to the membership relationship between the nodes and the elements, the material numbers corresponding to the elements where each node is located can be obtained, thus realizing the acquisition of the physical property parameters after the grid deformation.

[0058] In a specific implementation manner of this embodiment, step 104 includes the following sub-steps: Using the node coordinates and corresponding material numbers belonging to the elements of each initial grid, and fitting with an interpolation function to construct the obtained interpolation system as the parameter mapping model; Input the node coordinates of the deformed grid into the parameter mapping model and output the corresponding interpolation results; Take the material number closest to each interpolation result as the material number corresponding to each element of the deformed grid.

[0059] The interpolation system refers to the completed interpolation function.

[0060] In a more specific implementation manner of this embodiment, the interpolation function is a radial basis interpolation function.

[0061] It should be noted that in this embodiment, any available interpolation function can be used to fit and construct the parameter mapping model; in specific implementation, represent the interpolation function as , the function to be fitted is , based on the node coordinates and material numbers of each element of the given initial grid, by requiring to be equal to at the same coordinates, solve the system of equations to obtain the weight coefficients, thus completing the construction of the interpolation system and taking it as the parameter mapping model; after inputting the node coordinates of each element in the deformed grid into the parameter mapping model in sequence to obtain the corresponding interpolation results, take the material number close to the interpolation result as the material number corresponding to the element of the deformed grid; In a preferred implementation manner, consider using a function for interpolation operation with a radial basis function, that is, a radial basis interpolation function, to construct the parameter mapping model: ;

[0062] Require ;

[0063] Approximate function is a function at the th node coordinates is an interpolation function on, is a radial basis function, is the node index of the grid , is the total number of nodes in the grid, solve the above equations to obtain the weight coefficients , thus completing the construction of the parameter mapping model.

[0064] Exemplarily, taking the multi-physical parameter material of a three-dimensional test case as an example, such as Figure 12 shown, is the initial grid of the three-dimensional test case. The material number associated with the physical property parameters of Material 1 is 1, the material number associated with the physical property parameters of Material 2 is 2, and the material number associated with the physical property parameters of Material 3 is 3. The interpolation system is constructed in a radial basis function-based manner for testing. The test results are shown in Table 1 and Figure 13 shown; Table 1 Test Results

[0065] In this test case, if the calculated interpolation result is close to 1, the physical property parameters corresponding to this node are taken from Material 1. If the interpolation result is close to 2, the physical property parameters corresponding to this node are taken from Material 2, and so on.

[0066] In a specific implementation manner of this embodiment, step 104 includes the following sub-steps: Input the node coordinates of any unit of the initial grid into the initial artificial neural network model and output the predicted material number; Calculate the loss function value using the predicted material number and the corresponding material number, and iteratively optimize the initial artificial neural network model based on the loss function value to determine the parameter mapping model; Input the node coordinates of each node of the deformed grid into the parameter mapping model and output the material numbers corresponding to each unit of the deformed grid.

[0067] The initial artificial neural network model refers to the artificial neural network model to be trained.

[0068] The predicted material number refers to the material number output by predicting the material number for the grid node coordinates.

[0069] In a more specific implementation manner of this embodiment, the calculation process of the loss function value includes:

[0070] In the formula, is the loss function, is the total number of nodes in the grid, The node index of the grid , For the The node coordinates, is a variable model parameter, is the forecast material number, For the The material number corresponding to the node coordinates.

[0071] In a more specific implementation of this embodiment, the initial artificial neural network model is a fully connected neural network model.

[0072] It should be noted that in this embodiment, any available artificial neural network model can be used to fit and construct a parameter mapping model; in specific implementation, the node coordinates associated with the units of the initial grid are input into the initial artificial neural network model, and are transmitted to the prediction output (forward propagation) through various operations of each neuron to obtain the corresponding prediction material number. Before training, it is exemplarily known that The training purpose is to make the input-output relationship of the initial artificial neural network model close to these known data after it is trained. Therefore, the predicted material number is compared with the material number associated with the expected initial grid unit. The loss function value is calculated by the loss function to evaluate the difference between the two. The goal is to minimize the loss function value. The partial derivative of the loss function with respect to the weights in each neuron is calculated by the chain rule. The weights are updated (backward propagation) using the stochastic gradient descent algorithm. One forward propagation and back propagation are collectively called an epoch. When the loss function value obtained is less than the loss setting value or the number of epochs reaches the setting value, the model training is stopped, and the trained initial artificial neural network model is used as the parameter mapping model. The node coordinates of each unit of the deformed grid are input into the parameter mapping model to obtain the material number corresponding to each unit of the deformed grid. In a preferred implementation, a fully connected neural network model is considered to be used to construct a parameter mapping model; Figure 14 This is a schematic diagram of a fully connected neural network model, which includes three main parts: input layer, hidden layer, and output layer. Each layer usually contains multiple neurons, which are composed of input variables, weighted parameters, and activation functions. The parameters of the input layer are output to the first hidden layer. After each neuron in the hidden layer receives its input, the neuron model uses weight aggregation to process the data and outputs the data to the next layer under the action of the activation function. Figure 15It is a schematic diagram of the calculation process of a single neuron. Each neuron in each hidden layer calculates and outputs to the next layer according to its own model. This information transmission mechanism between layers is carried out layer by layer until the data reaches the output layer and produces the final output result. Among them, the data input to the neuron in the calculation process of each neuron is as follows: ;

[0073] In the formula, is the input from other neurons, is the output result obtained after the neuron calculation, is the activation function, is the th input value in the input from other neurons, is the total number of input values from other neurons, is the model weight, is the model bias, and constitute a set of variable model parameters in each neuron when training the neural network ; In specific implementation, the activation function can take the Sigmoid activation function: ; In the formula, is the input of the activation function.

[0074] This embodiment also provides a schematic flow chart for correctly obtaining physical property parameters of a multi-physical property parameter material under grid deformation conditions by using a mapping method, as shown in Figure 16 : 1) Establish its geometric model according to the size of the multi-physical property parameter material; 2) Perform grid division on the multi-physical property parameter material with the established geometric model to obtain the initial grid; 3) Construct a parameter mapping model with the node coordinates of the elements in the initial grid and the corresponding material numbers. Here, the interpolation function is used as an example to fit and construct the parameter mapping model. Correspondingly, the parameter mapping model is expressed as an interpolation system for interpolating material numbers; 4) Carry out phase change calculation and analyze the phase change situation of the multi-physical property parameter material; 5) Apply the grid deformation technology to deform the elements of the initial grid based on the phase change calculation results to obtain the deformed grid; 6) Substitute the node coordinates of each element of the deformed grid into the interpolation system to update and determine the corresponding material numbers, and apply the updated material number situation to the new grid configuration, that is, the deformed grid, to ensure that each grid element after deformation still has the correct physical property parameters.

[0075] It can be understood that, for the convenience and conciseness of description, only the general process of a method for obtaining physical property parameters during grid deformation is briefly described herein. The specific implementation process of each step can be understood by referring to the relevant content in the foregoing embodiments, and will not be elaborated herein.

[0076] In the embodiment of the present invention, a correspondence relationship between the node coordinates of the grid and the physical property parameters is established based on a mapping method, so as to obtain a parameter mapping model. After the grid is deformed, the physical property parameters of each unit of the deformed grid are obtained based on the parameter mapping model, ensuring the accuracy of the physical property parameters in complex phase change numerical simulations, especially in the case of multi-physical property parameter materials, providing more reliable support for engineering design and scientific research. At the same time, the operation is simple and has good versatility, without the need to repeat cumbersome data processing work after each deformation, avoiding the complexity and resource consumption of large-scale calculations, with high acquisition efficiency, and can be relatively easily implemented in the existing computing environment, having high practicality and promotion value, providing a more economical and convenient solution, and helping to improve the reliability of obtaining physical property parameters during grid deformation.

[0077] Please refer to Figure 17 , Figure 17 which is a structural block diagram of a device for obtaining physical property parameters during grid deformation provided by an embodiment of the present invention.

[0078] A device for obtaining physical property parameters during grid deformation provided in this embodiment includes: A geometric modeling module 1701, configured to establish a geometric model according to the size parameters of the multi-physical property parameter material; A grid division module 1702, configured to divide the geometric model into grids to determine an initial grid, and assign corresponding material numbers to the units of the initial grid according to the physical property parameters of the multi-physical property parameter material; A grid deformation module 1703, configured to process the initial grid by using a grid deformation method to obtain a deformed grid; A deformation parameter acquisition module 1704, configured to fit and determine a parameter mapping model based on the node coordinates and material numbers of the units of the initial grid, and determine the corresponding material numbers based on the node coordinates of the deformed grid through the parameter mapping model.

[0079] Further, the deformation parameter acquisition module 1704 is specifically configured to: Use the node coordinates and corresponding material numbers of each unit of the initial grid, and perform fitting by using an interpolation function to construct an interpolation system as the parameter mapping model; Input the node coordinates of the deformed grid into the parameter mapping model and output the corresponding interpolation results; Use the material number closest to each interpolation result as the material number corresponding to each unit of the deformed grid.

[0080] Further, the deformation parameter acquisition module 1704 is specifically configured to: Input the node coordinates of any unit of the initial grid into the initial artificial neural network model to output the predicted material number; Calculate the loss function value by using the predicted material number and the corresponding material number, and iteratively optimize the initial artificial neural network model based on the loss function value to determine the parameter mapping model; Input the node coordinates of each node of the deformed grid into the parameter mapping model to output the material numbers corresponding to each unit of the deformed grid.

[0081] Further, the calculation process of the loss function value includes: ; In the formula, is the loss function, is the total number of nodes of the grid, is the node index of the grid , is the th node coordinate, is the variable model parameter, is the predicted material number, is the th material number corresponding to the node coordinate.

[0082] Further, the interpolation function is a radial basis interpolation function.

[0083] Further, the initial artificial neural network model is a fully connected neural network model.

[0084] An embodiment of the present invention further provides a computer device, including a memory and a processor, wherein the memory stores a computer program; when the computer program is executed by the processor, the processor executes the steps of the method for obtaining physical property parameters during grid deformation according to any one of the above embodiments.

[0085] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program / instructions are stored, and when the computer program / instructions are executed by a processor, the steps of the method for obtaining physical property parameters during grid deformation according to any one of the above embodiments are implemented.

[0086] An embodiment of the present invention further provides a computer program product, including computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the method for obtaining physical property parameters during grid deformation according to any one of the above embodiments are implemented.

[0087] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0088] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical, mechanical or other forms.

[0089] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, in each embodiment of the present invention, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0091] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks or optical discs that can store program codes.

[0092] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for obtaining physical property parameters during grid deformation, characterized in that, Including: Establish a geometric model according to the dimensional parameters of the multi-physical property parameter material; Perform mesh division on the geometric model to determine the initial mesh, and assign corresponding material numbers to the elements of the initial mesh according to the physical property parameters of the multi-physical property parameter material; Use the mesh deformation method to process the initial mesh to obtain the deformed mesh; Based on the node coordinates and material numbers of the elements of the initial mesh, fit to determine the parameter mapping model, and determine the corresponding material numbers based on the node coordinates of the deformed mesh through the parameter mapping model.

2. The method for obtaining physical property parameters during grid deformation according to claim 1, wherein The step of determining the parameter mapping model based on the fitting of the node coordinates and material numbers of the elements of the initial mesh, and determining the corresponding material numbers based on the node coordinates of the deformed mesh through the parameter mapping model includes: Use the node coordinates and corresponding material numbers of each element of the initial mesh, and use the interpolation function for fitting, and construct the obtained interpolation system as the parameter mapping model; Input the node coordinates of the deformed mesh into the parameter mapping model, and output the corresponding interpolation results; Use the material number closest to each interpolation result as the material number corresponding to each element of the deformed mesh.

3. The method for obtaining physical property parameters during grid deformation according to claim 1, characterized in that, The step of determining the parameter mapping model based on the fitting of the node coordinates and material numbers of the elements of the initial mesh, and determining the corresponding material numbers based on the node coordinates of the deformed mesh through the parameter mapping model includes: Input the node coordinates of any element of the initial mesh into the initial artificial neural network model, and output the predicted material number; Calculate the loss function value using the predicted material number and the corresponding material number, and iteratively optimize the initial artificial neural network model based on the loss function value to determine the parameter mapping model; Input the node coordinates of each element of the deformed mesh into the parameter mapping model, and output the material numbers corresponding to each element of the deformed mesh.

4. The method for obtaining physical property parameters during grid deformation according to claim 3, wherein The calculation process of the loss function value includes: In the formula, is the loss function, is the total number of nodes in the grid, is the node index of the grid , is the th node coordinate, is the variable model parameter, is the predicted material number, is the th material number corresponding to the node coordinate.

5. The method for obtaining physical property parameters during grid deformation according to claim 2, wherein The interpolation function is a radial basis interpolation function.

6. The method for obtaining physical property parameters during grid deformation according to claim 3, characterized in that, The initial artificial neural network model is a fully connected neural network model.

7. An apparatus for obtaining physical property parameters during grid deformation, characterized in that, Including: A geometric modeling module for establishing a geometric model according to the dimensional parameters of the multi-physical property parameter material; A mesh division module for performing mesh division on the geometric model to determine the initial mesh, and assigning corresponding material numbers to the elements of the initial mesh according to the physical property parameters of the multi-physical property parameter material; A mesh deformation module for using the mesh deformation method to process the initial mesh to obtain the deformed mesh; A deformation parameter acquisition module for determining the parameter mapping model based on the fitting of the node coordinates and material numbers of the elements of the initial mesh, and determining the corresponding material numbers based on the node coordinates of the deformed mesh through the parameter mapping model.

8. A computer device, characterized in that, Including a memory and a processor, and a computer program is stored in the memory. When the computer program is executed by the processor, the processor executes the steps of the method for obtaining physical property parameters during mesh deformation according to any one of claims 1-6.

9. A computer-readable storage medium having computer programs / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, the steps of the method for obtaining physical property parameters during mesh deformation according to any one of claims 1-6 are implemented.

10. A computer program product, comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by a processor, the steps of the method for obtaining physical property parameters of grid deformation as described in any one of claims 1-6 are implemented.

Images