基于光滑支持向量机的机械振动时间序列分析

时间序列分析方法是动态系统建模的重要手段,传统的序列预测方法如统计和神经网络并不适用于复杂的非线性系统,为此引入了一种新的基于支持向量回归(SVR)的时间序列分析方法。为了降低计算的复杂度,采用了光滑化方法对SVR的基本算法进行改进,并应用于汽轮机振动数据序列,尝试建立汽轮机组振动状态模型。仿真结果表明:光滑支持向量回归(SSVR)算法具有良好的预测性能。与传统的时间序列预测方法(如神经网络)相比,SSVR算法具有更高的收敛速度和更好的拟合精度,有效地扩展了SVR的应用范围。

Time series analysis of mechanical vibration based on smooth support vector machine

Time series analysis is an important means of dynamic system modeling, but traditional series prediction methods such as statistics and neural network are not fit for complicated non-linear system.Therefore,a new method of time series prediction based on support vector regression(SVR) was introduced to resolve the problem of complicated non-linear system modeling.For the purpose of reducing calculation complexity,smooth arithmetic was imported to improve standard arithmetic of SVR.This new method was used to build vibrating model of turbine system.The results of simulation indicate that smooth support vector regression(SSVR) has excellent performance on time series prediction.Compared with traditional time series prediction method such as neural network,SSVR has faster convergence speed and higher fitting precision,which effectively extends the application of SVR.