Background
The carrier rocket flight process usually goes through some more complex physical processes, such as tail cover separation, boosting separation, interstage separation, star-rocket separation and other separation processes, longitudinal coupling vibration of the liquid rocket, and the like. The processes are often accompanied by the characteristics of short duration, strong multidisciplinary coupling and the like, are key focus links for current model development, and are important factors influencing the reliability of the carrier rocket, so that simulation recurrence and mechanism research work of related processes needs to be carried out by means of multi-professional comprehensive analysis, and the reliability of model development is fully improved.
A large amount of simulation work is developed in the development of aerospace products in China, a plurality of research results are formed, and some simulation application systems are constructed. The heterogeneous characteristics of models in different fields such as fluid, structure, dynamics and control are not fully considered by the simulation results or systems, so that the method obviously has defects in the aspects of model integration capacity, flow configuration capacity, combined simulation capacity and the like, and cannot be directly applied to coupled simulation application in related fields. At present, by means of secondary development and self-research of commercial software such as FLUENT, ABAQUS, ADAMS, MATLAB, LS-DYNA and the like, a plurality of heterogeneous models (constructed by different modeling languages of different commercial software) such as disciplinary professional software models of fluid, control, dynamics, load and the like are developed and used for mechanism analysis of ground or flight tests of the carrier rocket. However, these heterogeneous models are simulation analysis work developed from a certain angle for a specific specialty, and there are great differences between modeling tools and modes, non-uniform formats, and various operation configuration interfaces, which are not favorable for multi-specialty multi-factor comprehensive analysis. Meanwhile, in the running process of the simulation test, the access and the scheduling of various hardware resources such as a high-performance cluster computer, a graphic workstation and the like are involved. At present, for different simulation tests, a specific hardware application environment is usually selected and different interfaces and tools are developed, which results in waste of resources such as manpower and expenses.
In addition, each model is developed for a specific problem, and the model operation flow is solidified. When new simulation requirements are met, the utilization rate of a simulation test model is low due to a solidified operation process, so that repeated construction is brought, the problems of poor reusability and the like exist, the integration and the configuration of the simulation test of various heterogeneous models are not facilitated, and multi-professional interactive coupling simulation cannot be realized.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the simulation test system for the multidisciplinary heterogeneous model overcomes the defects of the prior art, unifies the interfaces of the heterogeneous models, can set a simulation flow according to the simulation purpose, can uniformly manage and configure simulation test resources, realizes multi-professional interactive coupling simulation, and improves the operation efficiency of the simulation test.
The technical solution of the invention is as follows: a multidisciplinary heterogeneous model-oriented simulation test system comprises a heterogeneous model modeling subsystem, a heterogeneous model packaging subsystem, a simulation test flow design subsystem, a simulation test resource allocation subsystem and a simulation test control subsystem;
the heterogeneous model modeling subsystem: receiving models in each field, and defining the operation flow of each model, interfaces between each model and a simulation test control subsystem before and after simulation calculation and interaction parameters according to the multi-professional coupling simulation purpose; the interface comprises a control interface and a parameter interface;
heterogeneous model encapsulation subsystem: configuring basic information of each model and resource information required by operation; generating a model parameter file of each model according to interaction parameters defined by the heterogeneous model modeling subsystem; outputting the basic information and the model parameter file of each model to a simulation test process design subsystem; outputting resource information required by the operation of each model to a simulation test resource configuration subsystem;
a simulation test process design subsystem: configuring a simulation test process according to the multi-professional joint simulation purpose, and outputting the configured simulation test process to a simulation test resource configuration subsystem and a simulation test control subsystem; setting a parameter matching relation of each interactive model in the simulation test process according to the basic information and the model parameter file of each model, and outputting the parameter matching relation to a simulation test control subsystem;
the simulation test resource allocation subsystem: configuring a computing node and an operation directory for each model in a simulation test flow according to resource information required by the operation of each model and currently available resource information of a simulation test system; the resource information refers to computing node information;
the simulation test control subsystem: according to interactive parameters defined by the heterogeneous model modeling subsystem, different test working conditions are defined for each simulation test process, under each working condition, an interface defined by the heterogeneous model modeling subsystem is utilized, according to the simulation test process, a model parameter file of each model and a parameter matching relation of each interactive model, a control instruction and an input parameter are sent to the corresponding model, a response signal and an output parameter are received, and sample data obtained in each joint simulation test is obtained according to the output parameter of each model.
The system also comprises a heterogeneous model interaction subsystem, wherein the heterogeneous model interaction subsystem is used for configuring the operation mode of each model to be single-step operation and defining an interface and interaction parameters between each model and the simulation test control subsystem in each calculation step;
the heterogeneous model encapsulation subsystem generates a process parameter file of each model according to interaction parameters defined by the heterogeneous model interaction subsystem;
when the multi-specialty joint simulation is coupling simulation, the simulation test control subsystem controls the single-step operation of each model through an interface under each calculation step of each model according to the simulation test flow and the process parameter file of each model, so that the cooperative control of the operation process of each heterogeneous model is realized, the joint simulation is completed according to the model parameter file of each model and the parameter matching relation of each interaction model, and sample data is obtained.
The model parameter file and the process parameter file respectively comprise parameter variable names, data types, attributes, default values, parameter units, parameter meanings, parameter types and dimension types, wherein the attributes refer to input or output, and the parameter types refer to before, in and after simulation calculation.
The method for configuring the computing node for each model by the simulation test resource configuration subsystem comprises the following steps:
(4.1) acquiring currently available computing node information of the simulation test system, and sequencing currently available computing nodes and node combinations from small to large according to the total number of cores to obtain a computing node sequence;
(4.2) extracting partial sequences with the total number of cores not less than x from the calculation node sequence as alternative sequences according to the number x of calculation cores required by the current model to be operated;
(4.3) extracting the computing node and the node combination with the minimum total number of cores from the alternative sequence, and if the computing node with the minimum total number of cores exists, directly distributing the model to the computing node with the minimum total number of cores; otherwise, selecting the node combination with the least number of computing nodes, and distributing the model to the node combination with the least number of computing nodes.
Compared with the prior art, the invention has the following beneficial effects:
(1) according to the invention, the interfaces of each heterogeneous model and the simulation test control subsystem are established through the heterogeneous model modeling subsystem and the heterogeneous model interaction subsystem, so that the unification of each heterogeneous model interface is realized, the simulation flow can be set according to the simulation purpose, and the multi-professional interactive coupling simulation is realized. Meanwhile, a simulation flow is set according to the simulation purpose, so that the simulation test integration and configuration of each heterogeneous model can be realized, the repeated construction is avoided, and the model reusability is improved.
(2) According to the invention, each model is configured to run in a single step through the heterogeneous model interaction subsystem, so that strict synchronization among the models during coupling simulation is ensured, and the cooperative running of various heterogeneous models is realized.
(3) The simulation test control subsystem realizes parameter matching of different variable names among the interactive models according to information such as parameter meanings, attributes and parameter types in the model parameter file and the process parameter file, and provides basic conditions for realizing joint simulation.
(4) The resource allocation subsystem for the simulation test allocates the computing nodes for each model according to the resource information required by the operation of each model, the currently available resource information of the simulation test system and the simulation test flow, so that the unified management and flexible allocation of computing resources are realized, and the operation efficiency of the simulation test is improved.
Detailed Description
The invention provides a multidisciplinary heterogeneous model-oriented simulation test system, which comprises a heterogeneous model modeling subsystem, a heterogeneous model interaction subsystem, a heterogeneous model encapsulation subsystem, a simulation test process design subsystem, a simulation test resource allocation subsystem and a simulation test control subsystem, as shown in figure 1. The multidisciplinary model includes a fluid computational model, a structural computational model, a control computational model, a load computational model, and the like.
Heterogeneous model modeling subsystem
Receiving models of each subject, and defining the operation flow of each model, interfaces (including control interfaces and parameter interfaces) between each model and a simulation test control subsystem before and after simulation calculation and interaction parameters according to the multi-professional coupling simulation purpose.
(II) heterogeneous model interaction subsystem
The heterogeneous model interaction subsystem is used for configuring the operation mode of each model to be single-step operation and defining an interface and interaction parameters between each model and the simulation test control subsystem in each calculation step. And the heterogeneous model packaging subsystem generates a process parameter file of each model according to the interaction parameters defined by the heterogeneous model interaction subsystem.
When the multi-specialty joint simulation is coupling simulation, the simulation test control subsystem controls the single-step operation of each model through an interface under each calculation step of each model according to the simulation test flow and the process parameter file of each model, so that the cooperative control of the operation process of each heterogeneous model is realized, the joint simulation is completed according to the model parameter file of each model and the parameter matching relation of each interaction model, and sample data is obtained.
The heterogeneous model interaction subsystem is mainly divided into the steps of model initialization, model input parameter receiving, model single-step calculation, model output parameter sending, model suspension and the like.
The specific operation process of model initialization is as follows: receiving the initialization parameters, performing initialization calculation and configuration of the model, generating an initial calculation model, and sending a model initialization completion command.
The specific operation process of the model single step calculation is as follows: and receiving a model operation command in the simulation test control subsystem, receiving a model input parameter, starting the model, performing the next calculation, solving and configuring, sending a model output parameter, and then sending a model single-step calculation finishing command.
(III) heterogeneous model encapsulation subsystem
Configuring basic information of each model and resource information required by operation; generating a model parameter file of each model according to interaction parameters defined by the heterogeneous model modeling subsystem, and generating a process parameter file of each model according to interaction parameters defined by the heterogeneous model interaction subsystem; outputting the basic information of each model, the model parameter file and the process parameter file to a simulation test process design subsystem; and outputting the resource information required by the operation of each model to a simulation test resource configuration subsystem.
The model parameter file and the process parameter file respectively comprise parameter variable names, data types, attributes, default values, parameter units, parameter meanings, parameter types and dimension types, wherein the attributes refer to input or output, and the parameter types refer to before, in and after simulation calculation.
Such as: parameter variable name: flow _ time; data type: a DOUBLE; the attributes are as follows: output; default values are: 0; unit of parameter: m/s; the meaning of the parameters: speed; the parameter types are as follows: before simulation calculation; dimension types are as follows: and (4) three-dimensional.
(IV) simulation test flow design subsystem
Configuring a simulation test flow for the simulation test control subsystem according to the multi-professional joint simulation purpose, and outputting the configured simulation test flow to the simulation test resource configuration subsystem and the simulation test control subsystem; and setting the parameter matching relation of each interactive model in the simulation test process according to the basic information and the model parameter file of each model, and outputting the parameter matching relation to the simulation test control subsystem.
The configuration of the simulation test flow by the simulation test flow design subsystem specifically comprises the steps of performing simulation flow editing, heterogeneous model association and simulation operation setting, wherein:
the specific operation process of the simulation flow editing and the association of the heterogeneous model is as follows: and determining each heterogeneous model according to the set simulation content, determining the operation sequence and parameter transmission relation of each heterogeneous model, and associating according to the operation sequence relation of the heterogeneous models.
The specific operation process of the simulation operation setting is as follows: and according to the time sequence relation of the simulation test operation, each heterogeneous model is set in a detailed mode, and the start-stop time, the simulation step length, the data synchronization point and the like are determined. According to the purpose of multi-professional joint simulation, the simulation test process comprises a serial process, a parallel process and synchronous operation configuration.
(V) simulation test resource allocation subsystem
And the simulation test resource allocation subsystem allocates a calculation node and an operation catalog for each model in the simulation test flow according to the resource information required by the operation of each model and the currently available resource information of the simulation test system.
The method for configuring the computing node for each model by the simulation test resource configuration subsystem comprises the following steps:
(4.1) acquiring currently available computing node information of the simulation test system, and sequencing currently available computing nodes and node combinations from small to large according to the total number of cores to obtain a computing node sequence;
(4.2) extracting partial sequences with the total number of cores not less than x from the calculation node sequence as alternative sequences according to the number x of calculation cores required by the current model to be operated;
(4.3) extracting the computing node and the node combination with the minimum total number of cores from the alternative sequence, and if the computing node with the minimum total number of cores exists, directly distributing the model to the computing node with the minimum total number of cores; otherwise, selecting the node combination with the least number of computing nodes, and distributing the model to the node combination with the least number of computing nodes.
The above process is exemplified as follows:
the coupling simulation test has two models, namely a model A and a model B; the resource information required by the model A and the model B is 32 cores.
When the simulation test control subsystem starts the model a, the simulation test resource configuration subsystem checks and confirms currently available computing node resources of the system, and if three types of computing node resources of the simulation test system are currently available, namely a 64-core node, a 32-core node and two 16-core nodes, the computing node sequence is as follows: a 64-core node, a 32-core node, two 16-core nodes forming a 32-core node, and two 16-core nodes. The alternative sequences extracted according to the method are as follows: a 64-core node, a 32-core node, and two 16-core nodes. Model a is assigned to one 32-core node.
When the simulation test control subsystem starts the model B, if the model A is finished running, the simulation test resource configuration subsystem rechecks and confirms currently available computing node resources of the system, and if three types of computing node resources of the simulation test system are available at present and are respectively a 64-core node, two 16-core nodes and four 8-core nodes, the computing node sequence is as follows: a 64-core node, a 32-core node formed by two 16-core nodes, a 32-core node formed by a 16-core node and two 8-core nodes, a 32-core node formed by four 8-core nodes, two 16-core nodes and four 8-core nodes. The alternative sequences extracted according to the method are as follows: a 64-core node, two 32-core nodes composed of 16-core nodes, and four 32-core nodes composed of 8-core nodes. Model B is assigned to a 32-core node combination of two 16-core nodes.
(VI) simulation test control subsystem
The simulation test control subsystem unifies the model parameter files of each heterogeneous model according to the purpose of the simulation test, generates simulation test parameters, and sets simulation test variables, wherein the specific operation process is as follows: according to the purpose of the simulation test, selecting a test working condition parameter variable from simulation test parameters, setting a value space of the parameter variable, selecting a method for generating a variable value sequence, generating an input sample of the simulation test parameters, and completing the configuration of the simulation test variables.
According to interaction parameters defined by a heterogeneous model modeling subsystem, different test working conditions are defined for each simulation test process, and the specific operation process of the multi-working condition setting is as follows: and according to the working condition of the simulation test, selecting one of methods such as a full factor method, an orthogonal method or a uniform design method by using the configured simulation test variable, generating various working condition combinations of the simulation test, and completing the multi-working condition setting of the test.
Under each working condition, an interface defined by the heterogeneous model modeling subsystem and an interface defined by the heterogeneous model interaction subsystem are utilized, control instructions and input parameters are sent to the corresponding models according to the simulation test process, the process parameter file of each model, the model parameter file of each model and the parameter matching relation of each interaction model, the cooperative control of the operation process of each heterogeneous model is realized, response signals and output parameters are received, and sample data obtained in each joint simulation test is obtained according to the output parameters of each model.
As shown in fig. 2, for a single-step operation flow chart of the heterogeneous model of the present invention, taking a model M as an example, firstly, simulation initialization is performed to complete interface and interaction parameter setting, then the model performs single-step iterative computation under the control of a simulation test control subsystem, if the computation step is finished, simulation post-processing is performed, otherwise, whether other models are called is determined, if so, the model output parameter is sent to a model N to be called to perform computation of the calling model N, the output parameter of the calling model N is output to the model M, and the above steps are continuously repeated. Until the model M operation is finished.
In fact, the simulation test control subsystem can also implement event-driven simulation, and when a predefined trigger event arrives, the current operation is interrupted, and the corresponding model is started. Fig. 3 is a flow chart of a simulation test run considering event driving, starting from a calculation model. Firstly, a calculation model A is calculated to obtain a calculation result, and then the following steps are executed:
if the time step has no downstream calculation model, judging whether the simulation of the current event is finished, if not, entering the simulation calculation of the next time step, if so, judging whether the next simulation event is present, if not, finishing the simulation, and if so, exciting the next simulation event and executing a new simulation time step;
if the time step has a downstream calculation model, the downstream calculation model starts to calculate to obtain a calculation result. And continuously and repeatedly executing the steps until the whole joint simulation is completed.
Those skilled in the art will appreciate that the invention may be practiced without these specific details.