SPH algorithm fragment power field simulation method based on LS-DYNA

Abstract

The invention discloses an SPH algorithm fragment power field simulation method based on LS-DYNA, and the method comprises the steps: obtaining the whole process information of shell SPH particles based on the SCL of LS-PREPOST; based on the space coordinates and influence radius information of the SPH particles, the space relevance between the SPH particles is determined by constructing a KD tree. SPH particles in each natural fragment are identified through a depth-first search algorithm, all SPH particles contained in each natural fragment are searched and counted in sequence, and then the mass, the volume, the component velocities in the X, Y and Z directions, the total velocity and the centroid coordinates of the fragments are calculated. And based on the identified SPH particle information of each natural fragment, performing morphological and dynamic reconstruction on each fragment to form a fragment model. Through the reconstruction model, the distribution condition of the natural fragments in different time and space can be simulated and predicted, and finally the full space-time fragment power field is generated. According to the method, the modeling process is simplified, and a feasible scheme is provided for large-scale fragment simulation in engineering application.

Description

Technical Field

[0001] The invention belongs to the technical field of numerical simulation, and in particular relates to a natural fragment spatial distribution simulation method based on an SPH algorithm of LS-DYNA. Background Art

[0002] The axisymmetric shell structure is a common structural form of natural fragmentation warheads. Driven by the detonation of internal explosives, this type of shell will expand, break and form random high-speed fragments. This process has long been the focus and difficulty of research in weapon damage effectiveness evaluation. At present, both at home and abroad, the traditional method of combining experiments with empirical formulas is still mainly used to evaluate the power field of natural fragments. An efficient and reliable method for simulating the power field of natural fragmentation warheads has not yet been formed.

[0003] In recent years, with the rapid development of numerical simulation technology, the simulation methods for the expansion and rupture of natural fragment warheads under internal explosions have gradually diversified, mainly including Lagrange algorithm, Euler algorithm, ALE (Arbitrary Lagrange Euler) algorithm and SPH (Smoothed Particle Hydrodynamics) algorithm. However, the traditional Lagrange algorithm, Euler algorithm and ALE algorithm have significant limitations when dealing with interface contact and large deformation problems under high-speed, instantaneous and high-intensity loading of materials. Especially in the process of fragment expansion and rupture, these algorithms are prone to grid distortion and cannot effectively capture the violent deformation and fragmentation of materials. In contrast, the SPH algorithm, as a meshless numerical method, uses discrete particles to describe macroscopic continuous media, and shows obvious advantages in dealing with instantaneous large deformation of materials. The algorithm avoids the problem of unit deletion caused by grid distortion, can ensure the conservation of mass, momentum and energy, and thus simulate the dynamic fragmentation process of materials more realistically. Therefore, the SPH algorithm is more in line with the actual situation and has significant advantages in the numerical simulation of the rupture and dispersion of natural fragment warheads.

[0004] The existing LSDYNA solver not only has a mature SPH algorithm, but also has a rich material library and high computational efficiency. However, its numerical model establishment and SPH particle conversion still need to be operated manually by clicking, and it is impossible to achieve automation and batch processing. Publication No. CN118246256A proposes a fragmentation spatial distribution simulation and verification method, which uses the LSDYNA solver to perform numerical simulation on the traditional anti-explosive warhead, and combines the Shapiro formula and the uniform distribution assumption to obtain a fragmentation spatial distribution. However, this method requires skilled LSDYNA pre-processing operation experience when establishing the numerical model of the warhead, including complex geometric modeling, material parameter definition, and SPH particle conversion operations, and it is not easy to model and set parameters for large-scale models.

[0005] At the same time, the post-processing program of LSDYNA has significant limitations, and it is impossible to effectively identify and count the key information of natural fragments generated by the SPH algorithm (such as spatial position, velocity, mass, etc.). This limitation has led to the limitation of the application analysis of the calculation results, especially the inability to quickly identify and output the detailed information of natural fragments, which greatly limits the application analysis of the calculation results. Publication No. CN103455669B proposes a quantitative automatic statistical method for fragments based on the calculation results of LSDYNA. This method reconstructs the fragment field through ANSYS based on the calculation results of LSDYNA, retrieves all unit information and outputs the mass, velocity and volume of each fragment. However, this method relies on the secondary input of ANSYS software and lacks statistics on the spatial coordinates of the fragments. Publication No. CN115099120A proposes a fragment identification method of the SPH algorithm of LSDYNA. This method constructs an SPH particle information matrix, screens and identifies all fragments, and finally outputs key information such as the mass and coordinates of each fragment. However, when the number of SPH particles reaches tens of thousands or even higher, the running memory required to construct the non-independent matrix that characterizes the interaction relationship between particles can reach hundreds of GB, which greatly increases the demand for computing resources and limits the practical application of this method. Summary of the invention

[0006] Aiming at the complexity and uncertainty of the temporal and spatial distribution of random fragments during the explosion of natural fragment warheads, the present invention proposes an innovative method that can efficiently and reliably simulate the temporal and spatial distribution of random fragments. This is achieved through the following specific technical solutions:

[0007] (i) Based on the SCL (Scripting Command Language) of LS-PREPOST, a natural fragmentation warhead model was parameterized and the SPH algorithm in LSDYNA was used to simulate the fragmentation process of the natural fragmentation warhead. Thus, the spatial coordinates, mass, velocity and impact radius of the shell SPH particles were obtained.

[0008] (ii) Based on the spatial coordinates and influence radius information of the SPH particles obtained in the previous step, the spatial correlation between the SPH particles can be quickly determined by constructing a KD tree (K-Dimensional Tree), that is, by determining which particles belong to the same natural fragment through neighborhood search.

[0009] (iii) The SPH particles in each natural fragment are identified by the depth-first search algorithm. All the SPH particles contained in each natural fragment are searched and counted in turn, and then the mass, volume, component velocity in the X, Y, and Z directions, total velocity, and center of mass coordinates of the fragment are calculated.

[0010] (iv) Based on the SPH particle information of each natural fragment identified in the previous step, each fragment is reconstructed in terms of morphology and dynamics to form a complete fragment model. Through this reconstructed model, the distribution of natural fragments in different time and space can be simulated and predicted, and finally a full-time and space fragment power field can be generated. This power field can intuitively show the destructive effect of fragments on surrounding targets, and combined with the vulnerability analysis of the target, it can provide a scientific basis for the destructive effectiveness evaluation of the warhead.

[0011] Compared with the prior art, the present invention has the following significant advantages:

[0012] 1. The present invention integrates geometric model definition, meshing, model conversion and parameter assignment through TUREGRID parametric modeling, LS-PREPOST SCL language, and SPH algorithm in LSDYNA, and realizes efficient modeling and high-precision simulation of the natural fragment warhead fragmentation process. The natural fragment spatiotemporal distribution generated by this method not only has high calculation accuracy, but also significantly simplifies the modeling process, providing a feasible solution for large-scale fragment simulation in engineering applications.

[0013] 2. By establishing a KD tree and a depth-first search algorithm, the present invention proposes a fast and accurate SPH particle swarm natural fragmentation identification method, which overcomes the limitations of the traditional LSDYNA post-processing software in statistics and identification of SPH fragmentation information. This method greatly simplifies the processing flow of SPH particle fragmentation information, ensures that a large amount of SPH particle data can be accurately processed in large-scale natural fragmentation simulations, solves the computational bottleneck of traditional identification methods when processing high-dimensional data, and ensures the accuracy and efficiency of identification.

[0014] 3. The SPH algorithm natural fragment reconstruction model method proposed in this invention can fully display the morphology and dynamic characteristics of natural fragments. By generating the full-time and space fragment force field, the damage effect of fragments on the target can be accurately simulated, thereby providing reliable technical support for the damage effectiveness of the warhead and the target vulnerability analysis. This model has broad application prospects in the military industry and can provide an important reference for warhead design and evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is the parametric modeling of natural fragment warhead and the SPH conversion display image.

[0016] Figure 2 It is a simulation of the fragmentation of a warhead caused by natural fragments.

[0017] Figure 3 It is an overall flow chart of the method of the present invention.

[0018] Figure 4 This is the image showing the SPH particle recognition results.

[0019] Figure 5 It is the SPH fragmentation force field. DETAILED DESCRIPTION

[0020] The present invention is described in detail below with reference to the accompanying drawings.

[0021] Example 1

[0022] Step 1: Define the geometric parameters of the natural fragmentation warhead, including the thickness r of the shell d , diameter d and length h, and the cylinder command of TRUEGRID is called through the PYTHON script to define the two parts (part) of the shell and the explosive respectively to generate the TRUEGRID input file (.tgf) to describe the geometry and mesh division of the finite element model. Next, the subprocess module of PYTHON is used to call TRUEGRID to execute the generated .tgf file to generate the finite element model. Subsequently, the SCL language is used in LSPREPOST to write an automated script to further process the generated finite element model. The SCL script will complete the following operations: First, the finite element mesh model is converted into an SPH model by converting the mesh units into SPH particles; second, the K file containing the material model definition and calculation parameters is imported into the SPH model through the keyword *INCLUDE. Finally, the generated command file (cfile) is executed by calling LSPREPOST through the PYTHON script to realize the conversion of the SPH model and the assignment of calculation parameters, and finally a complete numerical calculation model is established, such as Figure 1 The natural fragmentation warhead fragmentation simulation results are shown in Figure 2 shown.

[0023] Step 2: Use a PYTHON program to read the spatial coordinates, mass, velocity, volume, and influence radius information of SPH particles at each time step of the natural fragment warhead. Establish a KD tree based on the spatial coordinate information of the SPH particles, and perform a neighborhood search for each particle by looping through them. For the i-th particle, obtain the influence radius R i of this particle as the maximum search range, and then use the query_ball_point method of the KD tree to find all adjacent particles within the influence radius of this particle. For each found neighbor particle j, if the condition i < j is satisfied (to avoid duplicate records), then calculate the Euclidean distance D between the i-th particle and the j-th particle, and compare it with the larger value R = max(R i , R j ) of Ri and Rj. If D is less than R, it means that these two particles are related, and add the corresponding SPH numbers to the linked_particles list. Finally, the program loops through all SPH particles and outputs all found related SPH particle pairs one by one.

[0024] Step 3: Use the collections module in PYTHON to create a graph data structure, where the keys are the IDs of the SPH particles and the values are the sets of other SPH particles connected to them. Then, according to the given linked_particles list, loop through each pair of connected SPH particle numbers and add the relationship between each pair of SPH particles to the undirected graph. Then, use the depth-first search algorithm (DFS) to find all connected components, that is, all SPH particles contained in each natural fragment. The DFS function receives a starting node, a set of visited nodes, and a set of current connected components as parameters, and uses a stack to perform iterative search. The initial state of the stack is the incoming node. Each time a node is popped from the stack, if the node has not been visited, mark it as visited and add it to the set of current connected components. Next, add all unvisited neighbor nodes directly connected to the current node to the stack, and repeat this process until the stack is empty. At this time, the traversal of a connected component is completed. Loop through each node in the undirected graph. If the node has not been visited, use it as the starting point to perform a DFS operation once, and add the found connected components corresponding to each natural fragment to the fragments list. For each isolated SPH particle, add it as an independent connected component (i.e., a single natural fragment) to the fragments list to ensure that each SPH particle belongs to a certain natural fragment. Finally, loop through all nodes to group all SPH particles and output the SPH particle information corresponding to each natural fragment. The recognition results are as Figure 4 shown.

[0025] Step 4: Based on the fragments list obtained in the previous step, renumber the natural fragments. Natural fragment i is all the SPH particles contained in the i-th row in the fragments list. That is, the sum of the masses of all SPH particles in the fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, and the mass of all fragments in the fragment force field is counted.

[0026] The velocity of each natural fragment is determined by the following method: The velocity of the i-th natural fragment in the X direction is It is the weighted velocity sum of all SPH particles in the natural fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, M i represents the mass of the ith natural fragment, vx j represents the speed of the jth SPH particle in the X direction. Similarly, the speed of the i-th natural fragment in the Y direction is and the velocity in the Z direction is The combined velocity of the final i-th natural fragment is The speeds of all fragments in the fragment force field are counted.

[0027] The centroid coordinates of each natural fragment are determined by the following method: The X coordinate of the i-th natural fragment is It is the weighted X-coordinate sum of all SPH particles in the natural fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, M i represents the mass of the ith natural fragment, x j represents the X-coordinate of the j-th SPH particle. Similarly, the Y-coordinate of the i-th natural fragment is and the Z coordinate is The center of mass coordinates of all fragments in the fragment force field are counted.

[0028] Step 5: Based on the SPH particle information of each natural fragment obtained in step 3, the spatial coordinate information of all SPH particles containing the natural fragment i is extracted, and the minimum outer convex polyhedron of the particles is generated using the ConvexHull algorithm to reconstruct the geometric shape of the natural fragment. Then, the matplotlib module is used to visualize the constructed convex hull, so that the spatial distribution and morphology of the natural fragments can be intuitively displayed, as shown in the figure below. Figure 5 As shown, the accurate simulation of the natural fragment force field is achieved, providing a reliable basis for effectiveness evaluation and target vulnerability analysis.

Claims

1. A fragment force field simulation method based on the SPH algorithm of LS-DYNA, characterized by: The steps include: (i) Based on the SCL language of LS-PREPOST, a natural fragmentation warhead model was parameterized and the SPH algorithm in LSDYNA was used to simulate the fragmentation process of the natural fragmentation warhead; the spatial coordinates, mass, velocity and impact radius of the shell SPH particles were obtained; (ii) Based on the spatial coordinates and influence radius information of the SPH particles obtained in the previous step, the spatial correlation between the SPH particles is quickly determined by constructing a KD tree, that is, which particles belong to the same natural fragment through neighborhood search; (iii) The SPH particles in each natural fragment are identified by a depth-first search algorithm, and all SPH particles contained in each natural fragment are searched and counted in turn, and then the mass, volume, component velocity in the X, Y, and Z directions, total velocity, and center of mass coordinates of the fragment are calculated; (iv) Based on the SPH particle information of each identified natural fragment, each fragment is reconstructed in terms of morphology and dynamics to form a complete fragment model; through this reconstructed model, the distribution of natural fragments in different time and space can be simulated and predicted, and finally a full-time and space-time fragment power field can be generated; the full-time and space-time fragment power field can intuitively show the destructive effect of fragments on surrounding targets, and combined with the vulnerability analysis of the target, it provides a basis for the destructive effectiveness evaluation of the warhead.

2. The method for simulating the fragment force field using the SPH algorithm based on LS-DYNA according to claim 1 is characterized by: Define the geometric parameters of the natural fragmentation warhead, including the thickness r of the shell d , diameter d and length h, and define the shell and explosive parts part respectively through the PYTHON script by calling the cylinder command of TRUEGRID to generate the TRUEGRID input file .tgf to describe the geometry and meshing of the finite element model; then, use the subprocess module of PYTHON to call TRUEGRID to execute the generated .tgf file to generate the finite element model; then, use the SCL language to write an automation script in LSPREPOST to further process the generated finite element model. The SCL script will complete the following operations: first, convert the finite element mesh model into an SPH model by converting the mesh units into SPH particles; second, import the K file containing the material model definition and calculation parameters into the SPH model through the keyword *INCLUDE; finally, call LSPREPOST through the PYTHON script to execute the generated command file cfile to realize the conversion of the SPH model and the assignment of calculation parameters, and finally establish a complete numerical calculation model.

3. The method for simulating the fragment force field using the SPH algorithm based on LS-DYNA according to claim 2 is characterized by: Use a Python program to read the spatial coordinates, mass, velocity, volume, and influence radius information of SPH particles at each time step of a natural fragment warhead; establish a KD tree based on the spatial coordinate information of SPH particles, and perform a neighborhood search for each particle by looping through them; for the i-th particle, obtain the influence radius R of this particle i as the maximum search range, and then use the query_ball_point method of the KD tree to find all adjacent particles within the influence radius of this particle; for each found neighbor particle j, if the condition i < j is satisfied, calculate the Euclidean distance D between the i-th particle and the j-th particle, and compare it with the larger value R = max(R i , R j ) of Ri and Rj. If D is less than R, it means that these two particles are correlated, and add the corresponding SPH numbers to the linked_particles list; finally, the program loops through all SPH particles and outputs all found correlated SPH particle pairs one by one.

4. The method for simulating the fragment force field using the SPH algorithm based on LS-DYNA according to claim 3 is characterized by: The collections module of Python is used to create a graph data structure, in which the key is the ID of the SPH particle and the value is the set of other SPH particles connected to it. Then, according to the given linked_particles list, each pair of interconnected SPH particle numbers is traversed, and the relationship between each pair of SPH particles is added to the undirected graph. Then, the depth-first search algorithm DFS is used to find all connected components, that is, all SPH particles contained in each natural fragment. The DFS function receives a starting node, a set of visited nodes, and a set of current connected components as parameters, and uses a stack for iterative search. The initial state of the stack is the passed node. Each time a node is popped from the stack, if the node has not been visited, it is marked as has been visited, and add it to the current connected component set; next, add all neighbor nodes that are directly connected to the current node and have not been visited to the stack, repeat this process until the stack is empty, and the traversal of a connected component is completed; traverse each node in the undirected graph, if the node has not been visited, use it as the starting point to perform a DFS operation, and add the connected component corresponding to each natural fragment found to the fragments list; for each isolated SPH particle, add it as an independent connected component, that is, a single natural fragment, to the fragments list to ensure that each SPH particle belongs to a natural fragment; finally, traverse all nodes, realize the grouping of all SPH particles, and output the SPH particle information corresponding to each natural fragment.

5. The method for simulating the force field of fragments using the SPH algorithm based on LS-DYNA according to claim 4 is characterized by: Based on the fragments list, the natural fragments are renumbered. The natural fragment i is all the SPH particles contained in the i-th row in the fragments list; the mass of the i-th natural fragment It is the sum of the masses of all SPH particles in the fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, and the mass of all fragments in the fragment force field is counted; The velocity of each natural fragment is determined by the following method: The velocity of the i-th natural fragment in the X direction is It is the weighted velocity sum of all SPH particles in the natural fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, M i represents the mass of the ith natural fragment, vx j represents the speed of the jth SPH particle in the X direction; similarly, the speed of the i-th natural fragment in the Y direction is and the velocity in the Z direction is The combined velocity of the final i-th natural fragment is The speed of all fragments in the fragment force field is counted; The centroid coordinates of each natural fragment are determined by the following method: The X coordinate of the i-th natural fragment is It is the weighted X-coordinate sum of all SPH particles in the natural fragment, N represents the number of SPH particles in the fragment, j represents the jth SPH particle, M i represents the mass of the ith natural fragment, x j represents the X-coordinate of the j-th SPH particle; similarly, the Y-coordinate of the i-th natural fragment is and the Z coordinate is The center of mass coordinates of all fragments in the fragment force field are counted.

6. The method for simulating the force field of fragments using the SPH algorithm based on LS-DYNA according to claim 5, characterized in that: Based on the SPH particle information of each natural fragment, the spatial coordinate information of all SPH particles containing the natural fragment i is extracted, and the minimum outer convex polyhedron of the particles is generated using the ConvexHull algorithm, thereby reconstructing the geometric shape of the natural fragment; then the matplotlib module is used to visualize the constructed convex hull, so that the spatial distribution and morphology of the natural fragments can be displayed intuitively, realizing the accurate simulation of the force field of natural fragments, and providing a reliable basis for effectiveness evaluation and target vulnerability analysis.

Images