摘要
本文提出一种以高阶多项式逼近系统运行状态与稳定性指标间的映射进行电力系统动态安全评价的模式识别方法。这种方法的优点是，进行训练几乎无任何困难，且速度很快，易适用于多事故情况。应用加权最小二乘法求多项式系数可保证在稳定与不稳定边界附近识别结果的精度，应用递推最小二乘法可以提高学习的灵活性和效率。另外，本文所提出的利用导数信息扩充样本的方法可以提高样本获取的效率。
A new pattern recognition method forpower system dynarnic security assessment using highorder polynomial to approximate the mapping of sys-tem operation state to stability index is presented inthis paper. The characteristic of this method is easytraining, effective and suitable for multi-contingency.The application of weighted least square method tosolving polyomial factor would ensure the precision ofrecognition results colse to boundary of stability andinstabiility. The application of recurrence least squaremethod would improve the flexibility and efficiency oflearning. Moreover,the method to expand samplesusing the derivative information propossed in the papermay improve the efficiency of obtaining samples.
出处
《电网技术》
EI
CSCD
北大核心
1995年第1期13-16,共4页
Power System Technology
关键词
电力系统
动态安全评价
模式识别
Power system Dynamic securityassesment Pattern recognition
