摘要
汽轮机振动直接影响机组的安全,进行振动数据预测对机组的稳定运行具有重要意义。支持向量机是一种发展较好的常用振动预测方法,由于其输入特征对预测精度的影响较大,因此特征选择非常关键。提出了采用相空间重构理论对振动进行预测的方法,并与2种常规方法进行了比较,验证了基于相空间重构、支持向量机预测方法的优越性。
Turbine vibration has serious effect on the security of the unit. Vibration data prediction is important to its stable operation. Support vector machine is a common good development method of vibration prediction. However, the input characteristics have great effects on the accuracy of prediction and feature selection has been one common concern. This article uses phase space reconstruction theory to predict vibration. And with the comparison of two conventional methods, it verifies that the prediction method based on phase space reconstruction is superior,
出处
《东北电力技术》
2013年第9期40-43,共4页
Northeast Electric Power Technology
关键词
相空间重构
支持向量机
振动预测
Phase space reconstruction
Support vector machine
Vibration prediction
