Research On New Computing Model Of Electric Power System Based On Network Computing

Along with the implementation of Three-Gorges Project, power transmission from West China to East China, the interconnection among large area power networks are being formed. In such a large system, dealing with security and stability analysis, economical operation, and real-time control has become an extremely complicated computation issue, which need the real-time or super real-time computation to form accurate control strategy. The present sequential and parallel computing does not meet current requirements. In this paper, the network computing is applied to power system, and tries to solve the complicated computation of large-scale power system. The following work is researched.The architecture model of Power Grid is put forward. Based on existing resources and computing requirement of power system, the architecture model of Power Grid is put forward, and some primary problems and key technologies are researched. The hardware and software resources in power system are integrated and shared efficiently, and many shortages are remedied in the prevenient power monitor system.The mathematic model and computing method of power system which is fit to network computing is studied. The flow and transient stability mathematic model are built based on the network computing environment. Based on the different partition schemes of same computing problem, redundancy computing is put forward, and middle computing result is revised each other. The dynamic partition and stepwise optimization methods are researched in iterate computation. These revise methods may accelerate astringency and enhance veracity.The running model and estimating methods of multitasks based on network computing environment are researched. Many tasks run on the network computing platform of power system synchronously. To complete tasks efficiently by making the best of network resources, the running model of multitasks and implementing flow frame are studied based on network computing platform. The estimating parameters, called the total speedup, total efficiency, and finish time are put forward.The P2P computing model of power flow is put forward. This computing model implements the P2P distributed management, static and dynamic dispatch. The computing resources may be dispatched with the change of network environment and computing task. So the local concentricity and global distribution dispatch is implemented. The computing load balance becomes easies, and computing efficiency is enhanced. By using redundancy computing and dynamic transference method, the fault tolerant become more easily and reliably.The virtual cluster computing scheme of the transient stability computation is put forward. By integrating the present network computing and cluster technology, the virtual cluster computing technology comes into being, which may exert the advantage of network and cluster adequately. According to the computing architecture of tasks, this computing model may build the relevant virtual cluster. So the computing efficiency may be enhanced. By using static and dynamic resource obligating technology and dynamic dispatch, the computation break with the fault of transient stability may be solved efficiently. The network computing model will become one of the most cogent methods for great-scale power system computation. By researching the above problems, some theory foundation may be lay to solve real-time and complex computation of large-scale power system.