Network Simplification Studies For Active Splitting Control

System splitting is the last defense line protecting power systems from wide-area blackout. When a power system is suffered some serious disturbances and emergency control strategies fail to bring it back to normal, splitting control will suggest cut the original interconnected network into several isolated islands. After system splitting, although the power system is operating in an abnormal degraded state, customers are continuing to be served.Traditional splitting control strategies are often from a large enough event set obtained by an off-line calculation method. Compared with the traditional one, Active Splitting Control (abbr. ASC), which is aiming at the entire power system stability and security, is expected to monitor the system states in real time, determine splitting points online when islanding operation is inevitable, coordinate distributed out-of-step splitting equipments, and then achieve successful splitting timely. When a severer disturbance happens, active splitting control can separate asynchronous generator groups into different islands, and keep power balance on the sub-systems.The latest achievement about the Active Splitting Control is based on Graph Theory and Ordered Binary Decision Diagram (abbr. OBDD), using the scheme of"strategy searching and verifying", where all the lines in the power network are assumed to be allowed to trip off. Therefore Splitting Control, especially for a large-scale power system, is very complicated in general because a combinatorial explosion of strategy space happens. Therefore, it is necessary to simplify the power network for getting rid of many unfeasible splitting strategies and decreasing the scale of strategy space before the scheme is applied.This paper designs two efficient methods of network simplification for ASC to decrease the system scale to an acceptable level:Firstly, the paper presents a new method of network simplification based on electric distance. By analyzing the physical connection and electric distance of nodes, load condition of lines and topological characteristics of original power networks, a simplified system is formed according to the proposed method, which is much smaller in scale than the original one but keeps most of the static and dynamic characteristics of the original system.Secondly, another efficient method of network simplification is presented based on "principle of proximity", which can further reduce scales of large power systems than current existed methods while keeping as many feasible strategies as possible. Testing results on the IEEE 68-bus system and China north-east grid show that proposed method is feasible and effective for reducing system scale and speeding up strategy searching and verifying.Thirdly, the impacts of the granularity of node to the size of feasible strategy space are discussed in details. In the meanwhile, some merits and demerits of above two methods are summarized in paper finally.In addition, the author has built up the active splitting simulation platform. Many cases have been tested on the platform. Some conclusions have been drawn from the test results. New research directions have been pointed out based on the analysis of simulation cases.