摘要
为了避免在故障筛选与排序中忽略严重故障而导致大停电的发生,提出了一种比较精确的新的故障筛选和排序方法。以动态安全域的求解为基础,在故障筛选时运用解析法求动态安全域,进而求解失稳度。以故障的失稳度作为故障筛选的指标,快速选出严重的故障,形成故障筛选集。在故障排序时对于筛选集中的故障线路运用BP神经网络求解动态安全域,从而求解概率不安全指标,以故障的概率不安全指标作为故障排序的指标,得到精确的故障排序。通过IEEE10机39节点系统算例验证了该方法能够快速、全面、准确地实现故障的筛选与排序,同时通过IEEE4机11节点系统验证了BP神经网络求解动态安全域的可行性,误差为0.0608,满足要求。
In order to avoid ignoring the serious faults in the screening and ranking which results in large-scale blackout of power grid, a new contingency screening and ranking method is proposed. It is based on solving the dynamic security domains, and can calculate dynamic security region of power system by analytic method to calculate the instability degrees. The instability degree of the fault is the indicator of contingency screening to quickly select the severe faults to form contingency screening set. Dynamic security region of power system can be calculated by BP neural network in contingency ranking to get probabilistic insecurity index. The probabilistic insecurity index of the faults is the indicator of contingency ranking to accurately rank. Case studies on New England 10-machine 39-bus system show that the proposed method can realize the contingency screening and ranking quickly, fully and accurately. Meanwhile, case studies on New England 4-machine 11-bus system show that the calculation of dynamic security region about BP neural network is feasible, the error is 0.0608, which can meet the requirements.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2016年第18期75-80,共6页
Power System Protection and Control
关键词
动态安全域
故障筛选
失稳度
故障排序
BP神经网络
概率不安全指标
dynamic security region
contingency screening
instability degrees
contingency ranking
BP neural network
probabilistic insecurity index
