Power System Distributed Training System Research

The control objective of power system is to provide safe and stable supply of electricity which must meet certain quality requirements. Along with the growing demand, power systems become increasingly complex. Thus it is necessary to develop and deploy training simulators for dispatchers of control centers and substation operators and regularly carry out joint anti-accident training to improve the ability of operators to deal with abnormal operating conditions. These simulators should support networked simulation for transmission operations, therefore, the network modeling and simulation process have to be coordinated hierarchically across the individual utilities.The most ideal result of the modeling of power network is to build a complete and exclusive digital model in computer system corresponds to a real phsical power network. As power system is a very complex system which is a mechanism-electric-magnetism coupling system and made up of large numbers of devices, the modeling of power network is processed by levels of utilities according to the hierarchical and regionalization dispatch principle. So, in fact, the digital model of power network is a distributed and hierarchical model. In a distributed training proecess, the simulation computing processes of these models is coupling and interaction each other to represent the dynamic process of power network. This thesis focuses on the application model which integrates distributed training simulator based on different operation platform from various vendors. The main results of work and research in this paper are as follows:Firstly, with detailed analysis of running data of simulation, some methods of improving the convergency of power flow have been proposed, include dynamically identify of key bus, adaptive detection of load power measuments and adaptive adjustment of inject power of bus. Not like methods adopted by offline power flow algorithm, the methods of this paper can dynamic adjust running paraemeters of network in simulation process.Secondly, a general model of relaying protection which used to protect the whole power system is proposed. In this model, the protection operation characteristic can be modeling with a group of logical devices which are interrelated. The parameter of model can be stored in database, so a general algorithm could be developed to simulate various relaying protections. Thirdly, a distributed simulation architecture for transmission operations integrating hierarchical training simulators across wide network is presented. A platform neutral and scalable system architecture is provided by introducing high level architecture which is a modern distributed interaction simulation system architecture.Finally, the principle of decomposition and coordination computing used to coordinate simulation model between hierarchical simulators and the process to build distributed simulation system using HLA services is discussed in detail, the multi-scale modeling of power network and the match and merge of hierarchical model is also discussed, then, a reasonable and practical distributed algorithm for distributed training simulation system is proposed in this paper.The result of this thesis shows that decomposition and coordination computing and networked simulation model could be used to build generic, robust, portable and extensible distributed training simulator.