基于ANN和等值发电机模型的快速暂态稳定计算

作者：黄宇保,王建全* 日期：2010-06-28/span> 浏览：2882 查看PDF文档

基于ANN和等值发电机模型的快速暂态稳定计算

黄宇保,王建全*
（浙江大学 电气工程学院,浙江 杭州 310027）

摘要:暂态稳定分析对于电力系统运行具有重要的意义,针对暂态稳定时域仿真方法计算速度过慢的缺点,首先提出了应用于快速暂态计算的发电机参数等值方法,这种方法可以避免迭代解网络方程,能在保证计算精度的基础上显著减少暂态稳定计算时间,每个迭代步对发电机功角初值进行预测后则能够进一步减少解网络方程次数。算例仿真证明,粒子群算法优化得到的等值参数和基于神经网络的预测功角,在不同的系统运行方式下,能显著减少解网络方程次数和判定系统所处的稳定状态。算法具有计算精度高和收敛性良好的特点,功角预测和等值参数则有望应用于不同规模的系统中。
关键词:快速暂态稳定；参数等值；粒子群优化算法；神经网络；功角预测
中图分类号:TH7；TM74文献标识码:A文章编号:1001-4551(2010)06-0078-05

Research on fast transient stability of ANN and equivalent generator model

HUANG Yu-bao, WANG Jian-quan
(College of Electrical Engineering, Zhejiang University, Hangzhou 310027, China)

Abstract: It is well known that transient stability evaluation method for power system security operation is of great significance. Numerical integration method has been used widely and its drawback is that computation speed is too slow, so improving the speed is the main difficulty in time-domain simulation. A generator parameter equivalent methodology which can be applied to fast transient stability analysis by solving network equation without iteration was firstly presented, in addition computation time of time-domain simulation would be reduced by power angle prediction. The test system taking into account different operation conditions shows that the iteration times of solving network equations can be decreased by generator optimized parameters computed by particle swarm optimization algorithm (PSO) and power angle prediction based on artificial neural networks (ANN). This study explains that optimized parameters and angle prediction are expected to apply to numerical simulation in various power systems.
Key words: fast transient stability; parameter equivalent; particle swarm optimization(PSO); artificial neural networks(ANN); angle prediction