With constant improvement of intermittent energy source, its agglomeration effect and stochastic volatility will greatly influence the power system operation. It’s an inexorable trend that probabilistic power flow calculation shifts from offline analysis to real-time operation online. Among all probabilistic trend algorithms, probability flow calculation based on cumulant method is the fastest one; it has better prospect of online application. However, it requires random variables independent to each other, so the single slack bus always undertakes all unbalance power in conventional cumulant methods. When the fluctuation of system power injection becomes larger and larger, this calculation model will hardly adapted to the actual demand of the grid in the future. To improve the practicability, a improvement of probabilistic load flow based on cumulant method is proposed in this paper. By modify the calculation of the sensitivity matrix, the distribution of system unbalance power can be considered. The accuracy of this method is verified by the simulated analysis of the standard examples and actual power system model.

No abstract

With the construction of UHV power grid, integration of large-scale renewable clean energy, and large-scale energy base putting into operation, the power grid dispatching faced with more and more complex challenges. On the basis of existing research results, architecture of intelligent dispatching based on situation awareness is proposed, so as to accurately achieve prevention and control of the power system. The shortcomings of traditional dispatching mode are analyzed firstly, and the concepts and characterization approaches of grid situational awareness and operation state trajectory of power grid are then introduced. The overall objective of intelligent dispatching is presented, including data processing and integrated knowledge mining, predictive perception of grid operation, risk analysis and comprehensive early warning, so as to achieve "automatic cruise under normal operating conditions, automatic navigation under abnormal operating conditions ". The functional framework of intelligent dispatching is also proposed in details, including four major aspects of the perception and forecasts, risk analysis, decision-making support, and automatic control, as well as three supporting functions such as post-assessment of dispatching, trajectory index calculation, and human-computer interaction (HCI).Technical innovations to support automatic intelligent dispatching are discussed and organised in three levels, i.e. perception, comprehension and projection. The breakthroughs are: construction of index system, trajectory recognition based on massive information and knowledge mining, trajectory projection taking into accounts the uncertainties, online risk assessment and early warning, power grid intelligent decision-making support, automatic coordination of grid operation control, online assessment, natural human-computer interaction mode, and etc... These are the future research areas of automatic intelligent dispatching.

In this paper, a calculation method of static security analysis based on section topology relation is proposed. When topological structure of current data section is the same as that of the basecase section, it fully utilizes the factor tables of Jacobian matrix, and inverse matrix information of the basecase data section in rapid filtration of DC power flow. On the contrary, If topology changes compared with the basecase section, it adopts substation partial topological analysis technology as well as the method of partial factor table correction to increase the speed of calculation. Case studies with a practical power system indicate that the proposed method is correct and reasonable.

With the UHV (Ultra High Voltage) power grid construction and the interconnection of regional power grids, the scale of power grids in China is increasing rapidly. At the same time, significant uncertainty and variability is being introduced into power grid operation with the integration of large-scale renewable energy in power systems. All of these pose an enormous challenge to the operation control of power systems in China. For a long time, online static security analysis, as an important part of EMS (Energy Management System), has been an effective tool for power grid operation. However, it is increasingly difficult for traditional static security analysis in serial computing mode to be online applied in bulk power grids in China. A new practical parallel approach for online static security analysis is put forward in this paper. A multithread parallelism is introduced into contingency screening, detailed contingency evaluation and decision support for reducing the execution time. By employing the multithread technology, the hardware resources of multi-processor/multi-core computer can be fully used and the program can be speeded up effectively. The performance of the parallel static security analysis is demonstrated by tests on two large-scale power systems. The test results show that the proposed method can be online applied in real bulk power grids.