中文摘要：

解列是电力系统在受到大干扰、系统完整性无法保持的情况下将系统分解成几个孤岛的控制。如何高效寻找满足特定约束集的解列断面，是当前主动解列问题的研究热点。但是随系统规模的增大，现有基于图论模型的策略搜索算法，不同程度上存在维数灾困难。该文提出基于弱联接理论的决策空间预筛方法将有效降低预决策空间规模，再利用有序二元决策图将使求解大规模系统的解列策略（集）成为可能。对IEEE-68系统的测试结果显示，方法可在不损失可行解的前提下，极大程度上降低决策空间规模。

英文摘要：

Power system splitting （also called as system islanding） is to split the interconnected power grid into several islands when the system is subject to severe disturbances and system integrity can not be hold, so that the continuity of power service on the islands can be sustained. Recent interests have bean concentrated on how to find out splitting strategies that satisfy certain constraints in an efficient way. But methods based on graph-model always collapse because computation complexity explodes too severely with the increase of the scale of power systems. Instead of taking all branches as candidates, this paper proposes a weak-connection-based method to form a much smaller candidate strategy space before any searching. Thereafter, an OBDD （ordered binary decision diagram）-based method is applied to search feasible strategies within the candidate subspace. Tests on IEEE-68 systems show that the proposed method works well in large-scale power system.