| 引用本文: | 贾勇,何正友,廖凯.基于广域时空随机响应的低频振荡模态辨识[J].电力自动化设备,2016,36(12): |
| JIA Yong,HE Zhengyou,LIAO Kai.Low-frequency oscillation mode identification based on wide-area spatio-temporal stochastic responses[J].Electric Power Automation Equipment,2016,36(12): |
|
| 摘要: |
| 针对利用单通道信号无法准确辨识多个振荡模态，且不能估计振荡振型的问题，提出一种基于广域时空随机响应的振荡模态辨识方法。讨论了广域时空随机响应的向量自回归模型与系统振荡模态之间的联系，采用QR分解实现向量自回归模型参数的最小二乘估计；计算出振荡模态的模式参数，并通过系统随机响应的功率谱峰值确定系统的主导模态；通过新英格兰系统的蒙特卡罗仿真对模态辨识方法进行测试分析，结果表明利用广域时空随机响应能同时准确估计多个主导振荡模态的模式参数和振荡振型，与子空间辨识方法相比所提方法计算更简单有效；最后利用WECC系统的实测信号验证了所提模态辨识方法的适应性。 |
| 关键词: 电力系统 低频振荡 模态辨识 随机响应 向量自回归模型 |
| DOI：10.16081/j.issn.1006-6047.2016.12.008 |
| 分类号: |
| 基金项目:国家高技术研究发展计划(863计划)资助项目(2012-AA050208)；国家自然科学基金青年基金资助项目(51307146) |
|
| Low-frequency oscillation mode identification based on wide-area spatio-temporal stochastic responses |
|
JIA Yong, HE Zhengyou, LIAO Kai
|
|
School of Electrical Engineering, Southwest Jiaotong University, Chengdu 610031, China
|
| Abstract: |
| Since the multiple oscillation modes cannot be accurately identified only by the single-channel signals and the associated modal shapes cannot be estimated either, a method of oscillation mode identification based on wide-area spatio-temporal stochastic responses is proposed. The relationship between VAR(Vector AutoRegressive) model of wide-area spatio-temporal stochastic responses and system oscillation modes is discussed and the QR decomposition is applied to realize the least square estimation of VAR model parameters. The parameters of oscillation modes are calculated and the dominant mode of system is determined according to the power spectrum peak value of system stochastic responses. The proposed method is tested by the Monte Carlo simulation for New England system and results show that, the wide-area spatio-temporal stochastic responses can be used to estimate the modal parameters and modal shapes of multiple dominant oscillation modes accurately; and the proposed method is simpler and more efficient than the subspace identification method. The measured WECC system signals are applied to verify the flexibility of the proposed method. |
| Key words: electric power systems low-frequency oscillation mode identification stochastic response vector autoregressive model |
