Multidisciplinary coupling numerical simulation generality algorithm library and parallel computing method

Abstract

The invention provides a multidisciplinary coupling numerical simulation generality algorithm library and a parallel computing method, and relates to the field of aviation multidisciplinary coupling numerical simulation. Comprising a parallel coupling format library, a data mapping library, a grid support library, a data bus library, an MPI parallel library and a subject calculation interface library which are used for providing common algorithm support in multidisciplinary coupling simulation. According to the parallel computing method, a two-layer parallel architecture based on multiple plug-ins, multiple data and MPI is adopted, parallel in subjects, parallel between subjects, distributed data exchange and cross-process communication are achieved, and the parallel efficiency and numerical precision of coupling simulation are improved. The method is suitable for pneumatic + X type multidisciplinary coupling simulation, supports flexible expansion of disciplinary types and number, and has good parallel expandability and coupling precision.

Description

Technical Field

[0001] This invention relates to the field of multidisciplinary coupled numerical simulation technology in aerospace, specifically to a common algorithm library and parallel computing method for multidisciplinary coupled numerical simulation. Background Technology

[0002] Advanced aircraft design involves numerous multidisciplinary coupled simulation requirements, typically including aerodynamic-structural, aerodynamic-control, aerodynamic-noise, aerodynamic-stealth, and aerodynamic-structural-control simulations. These requirements can be summarized as "aerodynamics + X," where X represents aerodynamic-related disciplines and systems. Furthermore, as integrated design capabilities advance, the "X" considered evolves, manifesting as changes in disciplines or systems and an increase in their number. This necessitates that multidisciplinary coupled simulation software adapt to these changes in "X" to address the evolving simulation needs. Currently, numerical simulation technologies in aerodynamics, structure, and stealth are continuously improving and are widely and deeply applied in aircraft design, resulting in many mature CFD and CSD software programs. Simultaneously, numerous multidisciplinary coupled simulation software programs have emerged, but these are often developed for specific "X" scenarios and lack the ability to adapt to changes in "X." Consequently, they cannot build new multidisciplinary coupled simulation capabilities by integrating or coupling existing disciplinary computing software or code resources. Some multidisciplinary coupled simulation platforms have also emerged, which can organize computing software from different disciplines in a black box manner to achieve multidisciplinary coupled simulation, solving the problem of replacing "X". However, due to the coupling method and organization method, there are problems such as limited coupled simulation accuracy and parallel computing efficiency.

[0003] In response to the growing demand for multidisciplinary coupled simulation of "aerodynamics + X", especially the continuously growing demand for "X" disciplines, there is an urgent need to develop a common algorithm library for multidisciplinary coupled numerical simulation that is scalable across disciplines. This library should be able to continuously integrate and couple new disciplines, while making full use of existing discipline computing code resources to quickly build multidisciplinary coupled simulation software, establish engineering-practical multidisciplinary coupled simulation capabilities, and achieve long-term sustainable development, thus providing technical support for the integrated optimization design of advanced aircraft in "aerodynamics + X". Summary of the Invention

[0004] To address the limitations in accuracy and parallel computing efficiency of high-fidelity physical models in large-scale multidisciplinary coupled numerical simulations of aerospace applications, this application provides a common algorithm library and parallel computing method for multidisciplinary coupled numerical simulations. By utilizing existing disciplinary computing code resources, it enables the rapid construction of multidisciplinary coupled simulation software and improves the parallel scalability and numerical accuracy of such software.

[0005] This application provides the following technical solution: a common algorithm library and parallel computing method for multidisciplinary coupled numerical simulation, comprising: the common algorithm library includes a parallel coupling format library, a data mapping library, a grid support library, a data bus library, an MPI parallel library, and a subject computing interface library; The method employs a two-layer parallel architecture based on multiple plug-ins, multiple data, and MPI, and includes the following steps: S1: Initialize the multidisciplinary coupled simulation, including: each discipline plugin independently reads the mesh file and performs mesh partitioning, establishes MPI communication domains and communication lists within and between disciplines based on the MPI parallel library, constructs a multidisciplinary data bus through the data bus library and completes the initialization of variables for each discipline; S2: Initiate the outer loop iteration and determine the outer iteration variables based on the type of the coupled problem; S3: Start the inner loop iteration and determine the inner iteration variables according to the coupling convergence requirements; S4: Each subject plugin executes single-step subject calculations in parallel within its respective MPI sub-communication domain by calling the solver interface encapsulated in the subject calculation interface library; S5: Determine whether the inner iteration has converged; if it has not converged, then through the data mapping library, perform bidirectional data mapping between disciplines under the established interdisciplinary communication relationship; S6: Based on the mapping results, call the grid support library to update the computational grid for the relevant disciplines; S7: Determine whether the outer iteration has converged; otherwise, return to step S2. S8: Output the calculation results data for each subject and perform post-processing.

[0006] According to one embodiment of this application, the parallel coupling format library is used to provide implementations of various interdisciplinary coupling modes, including parallel loose coupling, parallel tight coupling, parallel unidirectional coupling, serial loose coupling, and serial unidirectional coupling.

[0007] According to one embodiment of this application, the data mapping library is used to provide surface-to-surface data conservation mapping, as well as volume-to-volume and volume-to-surface data compatibility mapping.

[0008] According to one embodiment of this application, the mesh support library is used to provide mesh deformation and mesh nesting functions based on an elastic body model.

[0009] According to one embodiment of this application, the data bus library is used for data management in multidisciplinary coupling processes and supports arbitrary data types and capacities.

[0010] According to one embodiment of this application, the MPI parallel library is used to establish and manage communication domains and communication relationships in parallel computing, and to provide MPI-based communication functions.

[0011] According to one embodiment of this application, the subject computing interface library is used to standardize and encapsulate heterogeneous subject solvers. The encapsulated interfaces include: a parallel communication interface, a problem definition interface, a grid access interface, a linear solver interface, a nonlinear solver interface, and a solution access interface.

[0012] According to one embodiment of this application, in step S2, if the type of coupling problem is dynamic coupling, the external iteration variable is the number of physical time steps; if the type of coupling problem is static coupling, the external iteration variable is the number of system coupling iterations.

[0013] According to one embodiment of this application, in step S5, the bidirectional data mapping between disciplines is achieved by calling the surface-to-surface data conservation mapping interface in the data mapping library.

[0014] This invention addresses the design problem of large-scale multidisciplinary coupled numerical simulation software for aerospace applications, with aerodynamics as the core. It proposes a common algorithm library and a parallel computing method for multidisciplinary coupled numerical simulation, meeting the accuracy and parallel computing efficiency requirements of high-fidelity physical models in large-scale multidisciplinary coupled numerical simulations of aerospace applications. This provides an effective solution for improving the parallel scalability of coupled software. The technical effects achieved by this invention mainly include three aspects: 1) It proposes a set of universally applicable integration interface specifications for various disciplines, enabling rapid integration of solvers from different disciplines; 2) It expands to accommodate different discipline types and numbers, meaning that parallel efficiency is not affected by the type or number of coupled disciplines; 3) It achieves global load balancing, efficient cross-disciplinary and cross-process point-to-point distributed communication, and full-process distributed parallelism, exhibiting good parallel efficiency; 4) It is compatible with multiple coupling mechanisms, ensuring the numerical accuracy of multidisciplinary coupled simulations. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the common algorithm library structure for parallel computing of multidisciplinary coupled numerical simulation in aerospace in this invention; Figure 2 is a flowchart of the construction of fluid-structure interaction computing capability in an embodiment of the present invention; wherein, Figure 2(a) is loose coupling and Figure 2(b) is tight coupling; Figure 3 This is a schematic diagram of the fluid-structure interaction parallel computing method in an embodiment of the present invention. Detailed Implementation

[0017] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0018] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] like Figure 1 As shown, this embodiment of the invention provides a common algorithm library and parallel computing method for multidisciplinary coupled numerical simulation, including: the common algorithm library includes a parallel coupling format library, a data mapping library, a grid support library, a data bus library, an MPI parallel library, and a subject computing interface library; The method employs a two-layer parallel architecture based on multiple plug-ins, multiple data, and MPI, and includes the following steps: S1: Initialize the multidisciplinary coupled simulation, including: each discipline plugin independently reads the mesh file and performs mesh partitioning, establishes MPI communication domains and communication lists within and between disciplines based on the MPI parallel library, constructs a multidisciplinary data bus through the data bus library and completes the initialization of variables for each discipline; S2: Initiate the outer loop iteration and determine the outer iteration variables based on the type of the coupled problem; In step S2, if the type of coupling problem is dynamic coupling, the external iteration variable is the number of physical time steps; if the type of coupling problem is static coupling, the external iteration variable is the number of system coupling iterations.

[0020] S3: Start the inner loop iteration and determine the inner iteration variables according to the coupling convergence requirements; S4: Each subject plugin executes single-step subject calculations in parallel within its respective MPI sub-communication domain by calling the solver interface encapsulated in the subject calculation interface library; S5: Determine whether the inner iteration has converged; if it has not converged, then through the data mapping library, perform bidirectional data mapping between disciplines under the established interdisciplinary communication relationship; In step S5, the bidirectional data mapping between disciplines is achieved by calling the surface-to-surface data conservation mapping interface in the data mapping library.

[0021] S6: Based on the mapping results, call the grid support library to update the computational grid for the relevant disciplines; S7: Determine whether the outer iteration has converged; otherwise, return to step S2. S8: Output the calculation results data for each subject and perform post-processing.

[0022] The technical solution of this invention mainly comprises two parts: a common algorithm library and a parallel computing method. The common algorithm library addresses common algorithmic problems encountered in the coupling of solvers across different disciplines in multidisciplinary coupled simulation software. It primarily includes parallel coupling formats, data mapping libraries, mesh support libraries, data bus libraries, MPI parallel libraries, and discipline computing interface libraries. Specifically, the common algorithms extracted from each library are as follows: The parallel coupling format provides implementations of parallel loose coupling, parallel tight coupling, parallel unidirectional coupling, serial loose coupling, and serial unidirectional coupling between disciplines, supporting coupling between two or more disciplines; The data mapping library provides surface-to-surface data conservation mapping, volume-to-volume and volume-to-surface data compatibility mapping, and supports multiple coupling scenarios; The mesh support library provides elastic body mesh deformation and mesh nesting implementation, supporting complex boundary motions in aerospace multidisciplinary coupled simulations; The data bus library provides subject data management, which is not limited by data type and capacity, nor by coupled subjects; The MPI parallel library provides communication relationship establishment, MPI communication domain management, and MPI-based communication functions; The subject-specific computing interface library provides a set of general subject-specific solver encapsulation interfaces, including parallel communication, problem definition, grid access, linear solvers, nonlinear solvers, and solution access.

[0023] The parallel computing method described herein addresses the efficiency problem of parallel computing in multidisciplinary coupled simulation software, improving the software's parallel scalability. This invention proposes a two-layer parallel computing method based on Multi-Plug-in Multi-Data (MPMD) and MPI, specifically including: 1. Parallel computing within a discipline; 2. Parallel computing between two or more disciplines; 3. Parallel computing for discipline data exchange based on distributed data; 4. Point-to-point communication across disciplines and processes.

[0024] To simplify the description, the following uses fluid-structure interaction (CFD-CSD) coupled simulation as an example, but is not limited to this, and the implementation of the present invention will be described in detail. Figure 2 shows a flowchart of the process of building fluid-structure interaction computational capabilities using the present invention. Figure 3 This invention provides a parallel computing method for fluid-structure interaction (FSI) simulations. The embodiments of this invention offer a two-layer, efficient parallel computing method for multi-plugin, multi-data scenarios in FSI simulations, such as... Figure 3 As shown, the main steps include: Step 1: Multidisciplinary Coupled Simulation Initialization. The multidisciplinary coupled software is driven by the workflow. The fluid and structural discipline plugins independently read the flow field mesh file and the structural mesh file, respectively, and then call the ParMetis interface in the mesh support library to perform mesh partitioning. Figure 1 ), and based on the MPI parallel library ( Figure 1 Establish communication lists for interdisciplinary communication domains, communication lists within discipline plugin communication domains, and communication lists between different communication domains of discipline plugins (m2n) in one go, such as Figure 3 As shown; then, data preprocessing operations are performed on each process, and a multidisciplinary data bus is established by calling the multidisciplinary data bus library; finally, initial values ​​are assigned to the flow field variables and structural variables to complete the initialization of the fluid and structural discipline plug-ins and the initialization of the fluid-structure interaction simulation.

[0025] Step 2: Initiate the outer loop iteration. For dynamically coupled problems, i.e., dynamic aeroelastic calculations, the outer iteration mainly focuses on the advancement of physical time, and the iteration variable is the number of physical time steps; for statically coupled problems, i.e., static aeroelastic calculations, the outer iteration mainly focuses on the convergence of the coupling variables in the iteration process, and the iteration variable is the number of iterations for the entire system, as shown in the n-loop in Figure 2(a)-Figure 2(b).

[0026] Step 3: Start the inner loop iteration. For dynamically coupled problems, i.e., dynamic aeroelastic calculations, the inner iteration mainly focuses on the convergence of the coupling variables in the iteration process, and the iteration variable is the pre-set number of iteration steps; for statically coupled problems, i.e., static aeroelastic calculations, the iteration variable is 1, as shown in the p loop in Figure 2(a)-Figure 2(b).

[0027] Step 4: Subject Plugin Calculation. Based on the communication domain settings in the MPI parallel library (e.g., ... Figure 3 (The fluid plugin is a sub-communication domain A, and the solid plugin is a sub-communication domain B). In sub-communication domains A and B, single-step calculations of each discipline are executed in parallel and independently by calling the open interfaces of the fluid and structural discipline plugins, and the calculation variables are updated.

[0028] Step 5: Internal iteration convergence judgment. If the residual value of the coupling variable drops to the given level or the number of iterations is executed to the preset number of iterations, then jump to step 2; if not, jump to step 8, as shown in the n-loop in Figure 2(a)-Figure 2(b).

[0029] Step 6: Bidirectional data mapping between flow field and structure field. Based on the interdisciplinary and cross-process m2n communication relationship established in Step 1, in the sub-communication domain C, by calling... Figure 1 The surface-to-surface data conservation mapping function interface in the data mapping library (RBF plugin) enables bidirectional high-fidelity mapping of flow field and structure field data, such as... Figure 3As shown in the figure. The data transmitted from the flow field surface mesh to the structure field surface mesh is pressure and velocity; while the flow field surface mesh transmits displacement.

[0030] Step 7: Calculate mesh update. Based on the surface mesh displacement of the flow field, the flow field calculation mesh is updated by calling the mesh deformation function in the mesh support library.

[0031] Step 8: External iteration convergence check. If the external iteration variable reaches the predetermined number of physical time steps, the outer loop iterates, and then jumps to step 9; if not, it jumps to step 2. As shown in the p loop in Figure 2(a)-Figure 2(b).

[0032] Step 9: Post-processing. Based on the user-specified output information, post-process the flow field and structural field data to facilitate visualization and analysis of the calculation results. Finally, terminate the parallel environment.

[0033] This invention utilizes existing disciplinary computing code resources to enable the rapid construction of multidisciplinary coupled simulation software, thereby improving the parallel scalability and numerical accuracy of such software.

[0034] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A common algorithm library and parallel computing method for multidisciplinary coupled numerical simulation, characterized in that, include: The common algorithm library includes a parallel coupling format library, a data mapping library, a grid support library, a data bus library, an MPI parallel library, and a subject computing interface library; The method employs a two-layer parallel architecture based on multiple plug-ins, multiple data, and MPI, and includes the following steps: S1: Initialize the multidisciplinary coupled simulation, including: each discipline plugin independently reads the mesh file and performs mesh partitioning, establishes MPI communication domains and communication lists within and between disciplines based on the MPI parallel library, constructs a multidisciplinary data bus through the data bus library and completes the initialization of variables for each discipline; S2: Initiate the outer loop iteration and determine the outer iteration variables based on the type of the coupled problem; S3: Start the inner loop iteration and determine the inner iteration variables according to the coupling convergence requirements; S4: Each subject plugin executes single-step subject calculations in parallel within its respective MPI sub-communication domain by calling the solver interface encapsulated in the subject calculation interface library; S5: Determine whether the inner iteration has converged; if it has not converged, then through the data mapping library, perform bidirectional data mapping between disciplines under the established interdisciplinary communication relationship; S6: Based on the mapping results, call the grid support library to update the computational grid for the relevant disciplines; S7: Determine whether the outer iteration has converged; otherwise, return to step S2. S8: Output the calculation results data for each subject and perform post-processing.

2. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The parallel coupling format library is used to provide implementations of various interdisciplinary coupling modes, including parallel loose coupling, parallel tight coupling, parallel unidirectional coupling, serial loose coupling, and serial unidirectional coupling.

3. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The data mapping library is used to provide surface-to-surface data conservation mapping, as well as volume-to-volume and volume-to-surface data compatibility mapping.

4. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The mesh support library is used to provide mesh deformation and mesh nesting functions based on the elastic body model.

5. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The data bus library is used for data management in multidisciplinary coupling processes and supports arbitrary data types and capacities.

6. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The MPI parallel library is used to establish and manage communication domains and communication relationships in parallel computing, and to provide MPI-based communication functions.

7. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, The subject-specific computing interface library is used to standardize and encapsulate heterogeneous subject-specific solvers. The encapsulated interfaces include: parallel communication interface, problem definition interface, grid access interface, linear solver interface, nonlinear solver interface, and solution access interface.

8. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, In step S2, if the type of coupling problem is dynamic coupling, the external iteration variable is the number of physical time steps; if the type of coupling problem is static coupling, the external iteration variable is the number of system coupling iterations.

9. The multidisciplinary coupled numerical simulation common algorithm library and parallel computing method according to claim 1, characterized in that, In step S5, the bidirectional data mapping between disciplines is achieved by calling the surface-to-surface data conservation mapping interface in the data mapping library.

Images