Quantification Study On Stability Of Dynamic Process After Controlled Islanding

With the wide-area measurements and high-speed communications technology used widely in the power system, stability and security control measures of power system are being gradually upgraded. As the third defense line of stability control of power system, splitting protection from traditional out-of-step protection based on off-line simulation, determining instability on local states to controlled islanding including online analysis coherent, intelligent search solution, remote active splitting control. Since controlled splitting is online decision, it is impossible to rely on offline simulation results to analyze the stability island after splitting. Therefore, the study of assess stability of island online after different splitting scheme is very important. Many scholars contribute lots of research on this area, which involve different constraints such as transmission power limit of line, reactive power balance into the objective function. Splitting solutions with different constraints can improve the stability of the island after splitting, but there are also differences on stability performance among these solutions and few people have discussed yet. Then quantification of stability performance after different islanding solutions are done in this paper, to help make a decision about which solution will be used in practice.This paper introduces both time-domain simulation and EEAC direct method to analyzing the dynamic process after splitting, to study whether different splitting algorithm has impact on system stability and the degree of influence. On one hand, the quantized time-domain simulation results analysis helps analyze the law of different solutions affecting the stability to guide the introduction of the constraints in search process; on the other hand, EEAC method can be used to make integrated assessment when splitting solutions got and select the most stable result online.Firstly, this paper compares different strategies in theory preliminarily, based on the application of graph theory to establish OBDD, weak connection and spectral clustering algorithm. In terms of the objective function, three methods correspond to the net power flow between islands, the smallest island imbalance, the smallest degree of dynamic coupling. We establish the simulation model of IEEE118 node by PSS\E software, including not only the classical generator models, but also the sixth-order complex models. In the time-domain simulation analysis, the paper defined two phases of the dynamic behavior of the system after splitting, in which the first phase named the transient process mainly reflects the non-synchronous generator running, the second phase of the steady process considers frequency offset of synchronous generator operation.Studies have shown that, in the time domain simulation, the impact of different strategies for splitting on the first synchronization phase is similar. The simulation result is below:the time of one generator takes to restore stable is similar when different splitting scheme used, but frequency offset is different. So, if the splitting algorithm does not consider power balance, the formation of island with maximum steady-state frequency offset frequency offset are poor performance. And the smallest island unbalance may appear larger degree trend to change original power flow of the system, then the island will lose stability.In the direct method aspect, the paper comes up with stability margin calculation process after islanding, which helps to compare stable performance of different strategies, without computing accumulated energy before islanding. The method SEEAC based on second order model of generation can be used to quantify stability of two groups of splitting system. Studies have shown that different splitting strategy appears similar dynamic stability, which is consistent with the time-domain simulation results.