|
| 摘要: |
| 在不损失原始数据主要信息的前提下，利用主成分分析进行输入特征变量的选择。考虑到主成分分析在数据标准化和处理非线性问题方面存在的局限性，采用一种改进的主成分分析法，进行特征选择。同时，针对电力系统暂态稳定分析中影响稳定性的关键因素在向量空间具有一定相似性的特点，采用动态聚类的方法将数据集分成若干并行子集，进一步压缩数据输入空间的大小，提高运算速度和效果。最后，利用关联分类法对数据进行分类和预测。通过对3机9节点系统的仿真试算，验证了该方法的有效性。 |
| 关键词: 电力系统 暂态稳定评估 特征选择 主成分分析 |
| DOI： |
| 分类号:TM712 |
| 基金项目: |
|
| Improved principal component analysis to feature selection for transient stability assessment |
|
YU Zhi-hong GUO Zhi-zhong
|
| Abstract: |
| The principal component analysis is used to select the input features premising no original data lost.With the consideration of its limitations in data standardization and nonlinearity processing,an improved principal component analysis is presented.Since the feature vectors affecting power system transient stability are dramatically similar to each other in the input space,the dy-namic clustering analysis is applied to divide the data set into several parallel subsets for further reducing the dimensionality of input space and improving calculation speed and effact.The associa-tive classification is employed to classify the data and assess the transient stability.The simulative calculation of a3-machine9-bus system verifies its effectiveness. |
| Key words: power system,transient stability assessment,feature selection,principal component analysis |
