论文部分内容阅读
提出了一种以关键运行特征识别和稳定薄弱环节辨识为目标的智能稳定评估方案,基于改进遗传算法与k阶近邻法(k-NN)相结合的稳定特征提取算法,实现对稳态运行信息中稳定关键特征的识别。在新英格兰10机39母线系统的仿真测试表明,算法能有效提取出反映不同区域稳定水平的少量关键运行特征变量,该特征较好地反映了失稳模式信息。通过构造基于BP网络的临界切除时间(CCT)预测器进一步验证了特征提取的有效性。基于特征提取结果与扰动位置的关联分析,提出了特征重合度判别方法,实现了对电网的稳定分区。
Aiming at the identification of the key operational characteristics and the identification of the stable weak links, an intelligent stability assessment scheme is proposed. Based on the improved genetic algorithm and the k-th nearest neighbor (k-NN) algorithm, a stable feature extraction algorithm is proposed to realize steady state operation information In the stability of the key features of the identification. The simulation test of bus system of New England 10-machine 39 bus shows that the algorithm can effectively extract a few key operating characteristic variables that reflect the stability of different regions, which can well reflect the instability mode information. The validity of feature extraction is further verified by constructing a critical cut-off time (CCT) predictor based on BP network. Based on the correlation analysis between feature extraction results and perturbation positions, a method of distinguishing feature coincidence degree is proposed, which realizes the stable partition of the power grid.