基于FMI的热插拔式联合仿真平台设计与实现

发布时间：2018-03-22 06:33

本文选题：信息物理融合系统　切入点：联合仿真　出处：《华东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：信息物理融合系统(Cyber Physical Systems,CPS)是信息计算和物理过程紧密结合与协作的系统,包括计算部分和物理部分,具有很强的异构性。此外,CPS集离散的系统行为和连续的物理过程于一体,是典型的混成系统。CPS具有开放性,通常运行在开放、不确定的环境中,导致系统行为具有不确定性。综合以上CPS的特性,模拟、验证分析CPS系统的性能以及安全性具有一定的挑战。联合仿真能一定程度上解决异构系统的建模与仿真问题,并且现有技术,如Functional Mockup Interface(FMI),为各工具的联合仿真提供统一的标准规范,任意工具通过实现FMI接口便能参与联合仿真,对联合仿真的发展具有一定促进作用。但是,一方面,实现FMI标准并导出FMU从而参与联合仿真是一项繁杂的工作,这限制了支持FMI标准工具的数量;另一方面现有的联合仿真算法效率有待提高。首先,设计并实现了基于FMI的热插拔式联合仿真平台,该平台支持模型的三种耦合方式:FMU式耦合、工具式耦合、插件式耦合。其中,FMU式耦合为传统的耦合方式;工具式耦合旨在为任意建模工具支持基于FMI的联合仿真提供更加高效、快捷的实现方式,从而使得更多工具支持基于FMI的联合仿真;插件式耦合旨在为代码定制的模型支持基于FMI的联合仿真提供更加高效、快捷的实现方式。支持三种耦合方式使得热插拔式联合仿真平台具有强大的仿真能力。其次,分析了现有的联合仿真算法,并提出了基于FMI的热插拔式的联合仿真算法,通过动态地改变参与联合仿真模型的数量以及控制联合仿真过程中耦合变量的交换次数从而提高联合仿真的效率。在仿真模型的热插拔实现上,为保证各模型建模阶段的独立性,我们设计了两种插拔情形:被动式插拔以及主动式插拔。针对两种不同的插拔情形,我们分别提出了两种实现方式:基于字符串的主动式插拔实现以及基于配置的被动式插拔实现。最后,我们采用两种联合建模与仿真的方案分析了 CPS中的经典案例——智能温控系统,充分彰显了热插拔式联合仿真平台的优势,并对联合仿真算法改进前后的效率进行了实验分析,结果证明,本文提出的热插拔式联合仿真算法能在一定程度上提高联合仿真效率。
[Abstract]:The information physics fusion system (Cyber Physical Systems / CPS) is a system of information computing and physical process closely combining and cooperating, including computing part and physical part, which has strong heterogeneity. In addition, it integrates discrete system behavior and continuous physical process. Is a typical hybrid system. CPS is open, usually run in an open, uncertain environment, resulting in uncertainty of system behavior. It is a challenge to analyze the performance and security of CPS system. Joint simulation can solve the modeling and simulation problems of heterogeneous systems to some extent, and the existing technologies, For example, the Functional Mockup Interface provides a unified standard for the joint simulation of various tools. Any tool can participate in the joint simulation by implementing the FMI interface, which can promote the development of the joint simulation to some extent. However, on the one hand, It is a complicated task to implement FMI standard and export FMU to participate in joint simulation, which limits the number of tools supporting FMI standard. On the other hand, the efficiency of existing joint simulation algorithms needs to be improved. This paper designs and implements a hot-swapping joint simulation platform based on FMI. The platform supports three kinds of coupling modes of model:: FMU coupling, tool coupling, plug-in coupling, in which FM-FU coupling is the traditional coupling mode. The purpose of tool-coupling is to provide a more efficient and fast way to implement joint simulation based on FMI for any modeling tool, so that more tools can support joint simulation based on FMI. Plug-in coupling is designed to provide a more efficient and fast implementation for code custom-made models to support joint simulation based on FMI. This paper analyzes the existing joint simulation algorithms, and puts forward a joint simulation algorithm based on FMI. The efficiency of joint simulation can be improved by changing the number of participating joint simulation models and controlling the number of exchange of coupling variables in the process of joint simulation. In order to ensure the independence of each modeling stage, we designed two kinds of plugging cases: passive and active. We propose two implementation methods: active plug based on string and passive plug based on configuration. Finally, We use two kinds of joint modeling and simulation schemes to analyze the classic case of CPS, intelligent temperature control system, which fully demonstrates the advantages of the hot-swapping joint simulation platform, and analyzes the efficiency of the joint simulation algorithm before and after the improvement. The results show that the joint simulation algorithm proposed in this paper can improve the efficiency of joint simulation to some extent.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP29;TP391.9

【参考文献】