Integrated distributed Tokamak simulation model management method
The integrated distributed tokamak simulation model management method addresses the lack of model management in existing systems by using a common library and hierarchical permission system to enhance efficiency and reduce redundancy in tokamak simulation.
Patent Information
- Application Number
- CN202411274855.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-13
AI Technical Summary
Current tokamak simulation platforms lack an integrated model management solution for distributed modeling, leading to inefficient and redundant construction of simulation models across multiple subsystems.
An integrated distributed tokamak simulation model management method involving a common basic model library maintained by framework developers, three-level permission management, and a hierarchical model submission and synchronization process to ensure efficient integration and sharing of models.
Enhances modeling efficiency and reduces redundancy by providing a unified management system for tokamak simulation models, ensuring accurate classification and integration, and improving the quality and reliability of the models.
Smart Images
Figure CN119129267A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of fusion control simulation, and in particular to an integrated distributed tokamak simulation model management method. Background Art
[0002] For the distributed digital tokamak simulation system, in order to better manage the models on the distributed cluster and reuse common functional modules and simulation models to the greatest extent, an integrated distributed simulation model management method is needed to support researchers to easily integrate new modules into cluster nodes after developing them, and to manage and share them in a unified manner. This model will be more suitable for complex modeling of multi-systems such as tokamaks.
[0003] The distributed construction of the digital tokamak means that each subsystem department needs to conduct collaborative modeling and simulation. The simulation modules and models designed and developed need to be managed and shared among the team to the greatest extent possible to avoid reinventing the wheel and improve modeling efficiency. At the same time, the simulation model library of each system needs to be reasonably divided, managed and saved, so the model modules constructed by each subsystem need to be managed in an integrated manner. In order to maximize the construction efficiency and utilization of the model, an integrated management mode for the digital tokamak simulation model is required. On the one hand, it provides and shares the basic modules required by each subsystem, and on the other hand, it ensures the accurate classification and efficient integration of the model.
[0004] The current tokamak simulation platform is mainly used to solve the physical problems of a single process, which also leads to the simulation platform mainly serving the model construction and simulation of independent nodes, without considering that the construction of large-scale distributed multi-node simulation models requires the participation of multiple people and multiple subsystems. The introduction of large-scale simulation systems such as digital tokamaks makes distributed model construction and simulation methods essential. Further providing a model management solution for distributed modeling will avoid repeated construction of similar modules between members, standardize model data ports, and enable teams to share models and collaborate on modeling more efficiently and conveniently.
[0005] Therefore, the present application proposes an integrated distributed tokamak simulation model management method. Summary of the invention
[0006] The purpose of the present invention is to propose an integrated distributed tokamak simulation model management method in order to solve the problem that there is no model management solution for distributed modeling in the background technology.
[0007] The technical solution of the present invention is an integrated distributed tokamak simulation model management method, comprising:
[0008] Provide a common basic model library maintained by simulation system framework developers to support basic logical operations of the model;
[0009] A three-level permission management model is used to manage the customized model library in a hierarchical manner. Researchers develop and test subsystem models in their personal customized libraries. After completion, they submit them to the subsystem administrator for review and verification. Once approved, the general system administrator merges them into the general model library and synchronizes them to each distributed node.
[0010] Researchers conduct model customization and development, including the definition of model names, the definition of shapes when representing models, the definition of model ports, and the definition of model parameters;
[0011] Model definition and integration: define the name, shape, length and width of the model, as well as the model description information, the number, name and data type of the model's input and output ports, and the number, name and data type of the model's parameters. Then define the model's behavior logic before the simulation starts, during the simulation iteration, and at the end of the simulation to complete the basic construction of the model.
[0012] Model submission and synchronization: researchers attach path information, subsystem administrators check definition specifications, determine classifications and register, the system copies files, administrators conduct tests, and after passing, use git tools to upload and synchronize, and the overall model library administrator reviews.
[0013] Optionally, the basic model library includes a waveform configuration module, a signal processing module and a symbolic operation module.
[0014] Optionally, the shape of the model when represented is defined as a rectangle or a circle, and the initial size of the model is defined by defining its length and width.
[0015] Optionally, the port definition of the model includes port name and port data definition. The port name needs to clearly indicate the input or output signal name and signal unit, so the port name should be defined in the form of "signal unit_signal name". If the signal is a dimensionless signal, such as a ratio or a normalized value, it should be directly in the form of "_signal name" starting with an underscore.
[0016] Optionally, the parameter names of the model use camelCase naming, that is, except for the first word which is completely lowercase, the first letters of the remaining words are all uppercase, and there is no space between words.
[0017] Optionally, the parameter data types of the model include int type, fl oat type, double type, bool type and char type.
[0018] Optionally, the parameter data of the model includes any one of a numerical value, a matrix and a string type. The numerical value needs to define the basic data type to which it belongs, and the matrix needs to define its matrix type including three matrix types: int type, float type and double type.
[0019] Optionally, the definition and integration of the model mainly involves defining the name, shape, length and width of the model and model description information, defining the number, name and data type of the input and output ports of the model, and defining the number, name and data type of the parameters of the model, and then defining the behavior logic of the model before the simulation starts, during simulation iterations and at the end of the simulation to complete the basic construction of the model.
[0020] Optionally, a three-level authority management mode is used to hierarchically manage the customized model library of the tokamak, specifically including: researchers customize and develop tests for the corresponding subsystem models in the personal customized libraries of their respective nodes. When the model test is completed and can be used for the digital tokamak, it will be submitted to the subsystem administrator, who will review and verify it and merge it into the subsystem's functional customization library. The subsystem's functional customization library will be merged into the overall model library by the overall system administrator and synchronized to other distributed nodes to complete the distributed customization and integrated modeling of the model library.
[0021] Optionally, the submission and synchronization of the model are specifically as follows: when the researcher submits the model in the designated personal customization library to the subsystem administrator, the path information of the personal customization library where the customized model is located will be attached. When the subsystem administrator receives the submission application and agrees to review it, the administrator will first check the definition specification of the model, and then determine the subsystem category to which the model belongs based on the model description, and register the model information in the management file of the subsystem library. The system will copy the model file to the directory of the subsystem category based on the path information of the personal customization library where the model is located. Then, the subsystem administrator will perform basic testing of the model, directly put it into the model running environment to perform calculations, and check whether it will report an error. If the test fails, the integration operation will be rolled back. When the test passes, the system will use the git tool to upload and synchronize the model file and management information to the overall model library. The administrator of the overall model library will conduct a final review. If it is found that the model is similar to models in other subsystem libraries, the review will be rejected. Otherwise, the review will be passed to complete the submission of the customized model.
[0022] Compared with the prior art, the present invention has the following beneficial technical effects:
[0023] Maintained by the developers of the simulation system framework, it supports the basic logical calculations of the model, provides convenience for researchers, and reduces repeated development. The hierarchical management of the customized model library ensures the review and verification of the model, and improves the quality and reliability of the model.
[0024] The distributed construction of the digital tokamak means that various subsystem departments need to conduct collaborative modeling and simulation. The simulation modules and models designed and developed need to be managed and shared among the team to the greatest extent possible to avoid reinventing the wheel and thus improve modeling efficiency.
[0025] The present invention provides an integrated management mode for digital tokamak simulation models. On the one hand, it provides and shares the basic modules required by various subsystems, and on the other hand, it ensures the accurate classification and efficient integration of the models, thereby maximizing the construction efficiency and utilization rate of the models. It is more suitable for complex modeling of multiple systems such as tokamaks, and improves the applicability and flexibility of the simulation system. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is the schematic diagram of the distributed model management mode;
[0027] Figure 2 Flowchart for model definition and integration;
[0028] Figure 3 The following is a flowchart for the model submission. DETAILED DESCRIPTION
[0029] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific embodiments.
[0030] Example
[0031] There are two main aspects to consider in model management of distributed digital tokamak simulation systems: one is the definition and integration of the model, and the other is the management solution for the model basic library and customized library.
[0032] For the large and complex Tokamak simulation model, many basic simulation modules are needed in the construction process to form the basic operations and processing of the plasma model. Therefore, a general basic model library is needed to ensure the basic construction of the model. This basic library is maintained by the developers of the simulation system framework and supports the basic logical calculations of the model. For example, the waveform configuration module will be responsible for generating the signal waveform required by the model, the signal processing module will be responsible for merging, selecting, and multiplexing the signals, and the symbolic operation will be responsible for the operations between the signals. These modules need to be provided to all simulation modelers, so their management is mainly developed and tested by the developers of the simulation framework, and uploaded to the general model library for synchronization to all nodes. Provide a general basic model library: maintained by the developers of the simulation system framework, support the basic logical calculations of the model, provide convenience for researchers, and reduce repeated development.
[0033] The customized model library for Tokamak is developed and tested by researchers of each subsystem and managed by subsystem and overall system administrators. Since the distributed simulation system model is deployed on each distributed node, it is necessary to Figure 1 The three-level permission management mode shown in the figure is hierarchically managed. Researchers will customize and develop tests for the corresponding subsystem models in the personal customization library of their respective nodes. When the model test is completed and can be used for the digital tokamak, it will be submitted to the administrator of the subsystem, who will review and verify it and merge it into the functional customization library of the subsystem. The functional customization library of the subsystem will be merged into the overall model library by the general system administrator and synchronized to other distributed nodes, thereby completing the distributed customization and integrated modeling of the model library. Three-level permission management mode: The hierarchical management of the customized model library ensures the review and verification of the model and improves the quality and reliability of the model.
[0034] Researchers need to follow the system framework rules when customizing and developing models, as shown in Table 1.
[0035] Table 1 Model definition rules
[0036]
[0037]
[0038] When defining the model name, all the words in the name should be in lowercase letters. When the name consists of multiple words, they are connected by underscores "_". The shape of the model is defined as a rectangle and a circle, and the initial size of the model is defined by defining its length and width. The port definition of the model includes the port name and port data definition. The port name needs to clearly specify the input or output signal name and signal unit, so the port name should be defined in the form of "signal unit_signal name". If the signal is a dimensionless signal, such as a ratio or a normalized value, it is directly defined in the form of "_signal name" starting with an underscore. Generally, a port has at least one data, and in most cases there is one column of data, that is, the data is one column and one row or one column and multiple rows, so the port can directly treat the data as a column, that is, an array. When defining the port, it is necessary to define the data type of the data in the array, including int type, fl oat type, double type, bool type and char type. In addition, the model also needs to define parameters. Model parameter names use camelCase, that is, except for the first word which is completely lowercase, the first letter of the remaining words is all uppercase, and there is no space between words. Parameter data types are more diverse than port data types. Parameter data can be a numeric value, in which case you need to define its basic data type, or a matrix, in which case you need to define its matrix type, including int type, float type, and double type. It can also be a string type.
[0039] The definition and integration of the model mainly go through the following steps: Figure 2The process shown in the figure. When integrating a customized model, researchers must first define the name, shape, length and width of the model and the model description information according to the rules in the previous article, define the number, name and data type of the model's input and output ports, and define the number, name and data type of the model's parameters. Then, define the model's behavior logic before the simulation starts, during the simulation iteration, and at the end of the simulation. This completes the basic construction of the model. When integrating the model, you need to first write the model name and model path into the management file of the personal customization library to complete the registration of the model in the customization library. Then, put each definition file of the model into the personal customization library directory of the corresponding path to integrate the model in the personal customization library. Finally, put the integrated model into the model execution environment for verification and testing, which completes the definition and integration of the model. Follow the system framework rules: When customizing and developing models, researchers must follow the system framework rules, including the definition of the model name, the definition of the shape when the model is represented, the definition of the model port, and the definition of the model parameters, to ensure the consistency and standardization of the model. By defining the model's name, shape, length and width, as well as model description information, defining the number, name and data type of the model's input and output ports, and defining the number, name and data type of the model's parameters, and then defining the model's behavior logic before the start of simulation, during simulation iterations, and at the end of simulation, the basic construction of the model is completed, thereby improving the model's construction efficiency and utilization rate.
[0040] The submission and synchronization of the model mainly involves the transmission of model information and the review of management personnel. Figure 3 As shown in the figure, when researchers submit the model in the designated personal customization library to the subsystem administrator, the path information of the personal customization library where the customized model is located will be attached. When the subsystem administrator receives the submission application and agrees to the review, he will first check the definition specifications of the model, then determine the subsystem classification to which the model belongs based on the model description, and register the model information in the management file of the subsystem library. The system will copy the model file to the directory of the subsystem classification based on the path information of the personal customization library where the model is located. Next, the subsystem administrator will perform basic tests on the model, directly put it into the model running environment to perform calculations, and check whether it will report an error. If the test fails, the integration operation will be rolled back. When the test passes, the system will use the git tool to upload and synchronize the model file and management information to the general model library. The administrator of the general model library will conduct a final review. If it is found that the model is similar to the models in other subsystem libraries, the review will be rejected. Otherwise, the review will be passed, thus completing the submission of a customized model. The researcher attaches the path information, the subsystem administrator checks the definition specifications, determines the classification and registers it, the system copies the file, the administrator tests it, and after passing, it is uploaded and synchronized using the git tool. The administrator of the general model library reviews it, ensuring the accuracy and completeness of the model.
[0041] The distributed modeling of the present invention provides a model management solution that will avoid repeated construction of similar modules between members, standardize model data ports, and enable teams to share models and collaborate in modeling more efficiently and conveniently. The distributed construction of the digital tokamak means that each subsystem department needs to conduct collaborative modeling and simulation, and the simulation modules and models designed and developed by them need to be managed and shared among the team to the greatest extent possible to avoid repeated "reinventing the wheel" and thus improve modeling efficiency.
[0042] Reasonable division, management and preservation of model library: The simulation model library of each system needs to be reasonably divided, managed and preserved, so the model modules constructed by each subsystem need to be managed in an integrated manner.
[0043] It is worth noting that in order to maximize the efficiency of model construction and utilization, an integrated management mode of digital tokamak simulation models is needed, which can provide and share the basic modules required by each subsystem on the one hand, and ensure the accurate classification and efficient integration of the model on the other hand. This makes it more suitable for complex modeling of multiple systems such as tokamaks, and improves the applicability and flexibility of the simulation system.
[0044] The above specific embodiments are only several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.
Claims
1. An integrated distributed tokamak simulation model management method, characterized in that: include: Provide a common basic model library maintained by simulation system framework developers to support basic logical operations of the model; A three-level permission management model is used to manage the customized model library in a hierarchical manner. Researchers develop and test subsystem models in their personal customized libraries. After completion, they submit them to the subsystem administrator for review and verification. Once approved, the general system administrator merges them into the general model library and synchronizes them to each distributed node. Researchers conduct model customization and development, including the definition of model names, the definition of shapes when representing models, the definition of model ports, and the definition of model parameters; Model definition and integration: define the name, shape, length and width of the model, as well as the model description information, the number, name and data type of the model's input and output ports, and the number, name and data type of the model's parameters. Then define the model's behavior logic before the simulation starts, during the simulation iteration, and at the end of the simulation to complete the basic construction of the model. Model submission and synchronization: researchers attach path information, subsystem administrators check definition specifications, determine classifications and register, the system copies files, administrators conduct tests, and after passing, use git tools to upload and synchronize, and the overall model library administrator reviews.
2. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The basic model library includes a waveform configuration module, a signal processing module and a symbol operation module.
3. The integrated distributed tokamak simulation model management method according to claim 2, characterized in that: The shapes of the model when represented are defined as rectangle and circle, and the initial size of the model is defined by defining its length and width.
4. The integrated distributed tokamak simulation model management method according to claim 3, characterized in that: The port definition of the model includes port name and port data definition. The port name needs to clearly indicate the input or output signal name and signal unit, so the port name should be defined in the form of "signal unit_signal name". If the signal is a dimensionless signal, such as a ratio or a normalized value, it should be defined in the form of "_signal name" starting with an underscore.
5. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The parameter names of the model are named using camelCase, that is, except for the first word which is completely lowercase, the first letters of the remaining words are all capitalized, and there is no space between words.
6. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The parameter data types of the model include int type, float type, double type, bool type and char type.
7. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The parameter data of the model includes any one of the numerical, matrix and string types. The numerical value needs to define the basic data type to which it belongs, and the matrix needs to define its matrix type including three matrix types: int type, float type and double type.
8. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The definition and integration of the model mainly involves defining the name, shape, length and width of the model, as well as model description information, defining the number, name and data type of the model's input and output ports, and defining the number, name and data type of the model's parameters, and then defining the model's behavior logic before the start of simulation, during simulation iterations, and at the end of simulation to complete the basic construction of the model.
9. According to claim 1, the integrated distributed tokamak simulation model management method adopts a three-level authority management mode to hierarchically manage the tokamak customized model library, specifically comprising: Researchers customize and develop tests for the corresponding subsystem models in the personal customization libraries of their respective nodes. When the model test is completed and can be used for the digital tokamak, it will be submitted to the subsystem administrator, who will review and verify it and merge it into the subsystem's functional customization library. The subsystem's functional customization library will be merged into the overall model library by the overall system administrator and synchronized to other distributed nodes to complete the distributed customization and integrated modeling of the model library.
10. The integrated distributed tokamak simulation model management method according to claim 1, characterized in that: The submission and synchronization of the model are specifically as follows: when the researcher submits the model in the designated personal customization library to the subsystem administrator, the path information of the personal customization library where the customized model is located will be attached. When the subsystem administrator receives the submission application and agrees to review it, he will first check the definition specification of the model, and then determine the subsystem category to which the model belongs based on the model description, and register the model information in the management file of the subsystem library. The system will copy the model file to the directory of the subsystem category based on the path information of the personal customization library where the model is located. Then, the subsystem administrator will perform basic testing of the model, directly put it into the model running environment to perform calculations, and check whether it will report an error. If the test fails, the integration operation will be rolled back. When the test passes, the system will use the git tool to upload and synchronize the model file and management information to the overall model library. The overall model library administrator will conduct a final review. If it is found that the model is similar to models in other subsystem libraries, the review will be rejected. Otherwise, the review will be passed to complete the submission of the customized model.