Design And Implementation Of Modeling And Processing Tools For Hybrid Dynamic System

Dynamic system modeling is a class of model approach which uses math equation as the expression of the real physical system. Based on the system state changes over time, dynamic systems can be divided into continuous-time systems and discrete-time systems. In the continuous-time system, the system state changes over time continuously, generally through various forms of differential equations (as differential equations, differential-algebraic equations, etc.) to set up the mathematical model, The status of the trigger signal is also the evolution of continuous-time, continuous-time model of the simulation was mainly falls on the corresponding numerical differential equation solver on the research and development; In the discrete-time model, system state changes over time discretely, generally through a variety of discrete forms of treatment mechanisms (such as automata, etc.) to carry out modeling, the activation signal for driving state change is called event, therefore, also known as discrete-time systems for discrete event systems. Currently pure continuous-time systems and pure discrete-time systems research has a more sophisticated stage.The so-called hybrid systems, is both the continuous sub-systems and discrete sub-systems existence, and the interaction between the two. The studies of the two systems use different physical laws and mathematical methods, so the research for hybrid systems is great more difficult compared to the system as a pure system. Modeling of hybrid dynamic systems is the more complex expression of real physical system, which better adapts to the model and simulation of the multi-area and complex system. Hybrid systems modeling approach as the field of modeling and simulation of a major hot spot are near ten years, things are still a lot of theoretical problems to be overcome, and its specific application of the corresponding technology is also required to carry out continuous testing and verification.At present, the modeling and simulation tools have emerged, from the international scale; particularly MATLAB/Simulink tool is the most popular. This type of modeling tools represent the current mainstream of the model description. Say general, Simulink uses a block diagram approach of the model description, with the following features: each modeling element is a general expression of an independent subsystem, and the implementation of a specific function, known as system modeling; each modeling element has its own data input and output ports, and thus have a model specific data stream, known as causal modeling; each model function has a corresponding implementation of the specific algorithm, it is still process-based.For its powerful ability of Modeling and Simulation, simulink has a wide range of applications. However, the obtain of this capacity is highly dependent on the MATLAB/Simulink platform, which may be very suitable for some of the more purely theoretical, or technical studies, but for large-scale system simulation software, if highly dependent on This commercialization of generic modeling platforms, to ensure that the modeling capabilities, it also has many obvious shortcomings:First, if large-scale dynamic system simulation software system pays the job of modeling and simulation fully to this kind of commercial software to achieve, then the whole system will greatly enhance the development costs, which involved the development of cost-effective system problem.Second, even with the commercial software to do the main job, because of its high degree of versatility, a number of specialized areas for the special modeling behavior, is also not satisfied by Simulink directly.In this paper, one of the main study ideas is to design a simple and valid formal model semantic by the block diagram approach, and to develop a portable module with C++ standard library for model building and pre-processing, which will be integrated into a large-scale application of simulation software as a independent module component. Then through the S-Function mechanism from Simulink within the model information format conversion, the model designed by the special application platform will be compiled and simulation by Simulink tool.Further more, base on the existing formal model that we have designed, we continue to develop corresponding Compilation module. Because the complication process of model, most basic, is to finish the convert form the formal model implemented by the editor to the formal model that can be identified by the simulator. Therefore, the design and implementation of the compilation mechanisms, not only need to deal with the topology structure of the diagram created by editor, but also need to consider the full range of numerical methods used by simulator, so the development of an efficient compiler is such a difficult thing. We designed the current compiler dealing mainly with the topology mechanism of the model effectively, and the design process of the numerical solution mechanism for simulator needs for more in-depth research.In summary, this paper has done the work about expression of hybrid model and corresponding model processing technology, as follows:1) Design a kind of formal semantics for hybrid model, and design a series of special primitive basic models to support the formal design.2) Design and implement an editor system prototype, which is divided into two modules: the GUI module and the modeling module. The GUI module is implemented for interactive visual model editing and the sequence operation of model; of modeling module, the whole model diagram is constructed by a hierarchical tree structure, and the model file format uses a simple and efficient line-format approach.3) Design and implement the pre-processing technology for model diagram. The sub-system deal with a diagram from editor, after some necessary eliminate operation for the diagram, output a series of internal data structures as a internal model. A simple model can be simulated directly.4) For complex models, we must compile the diagram before we can simulate it, and what's more, a special hybrid simulator is also needed. The current compiler compiling module owns a series of compiling technology for signals, including the inheritance mechanism of signal, the critical classification algorithms of events, the optimization strategy for signal scheduling, and so on.5) From the point view of numerical integrator, the output result of compiler is a unified system model. In this paper, based on the system-level modeling view, we present a global model of the system as the output of compiler.