提出一种基于非线性自回归时间序列模型(gereral expression for linear and nonlinear auto-regressive mod-el,简称GNAR模型)的机械系统状态识别与故障诊断方法。利用采集系统工作过程中的特征信号建立GNAR模型;用主成分分析策略生成...提出一种基于非线性自回归时间序列模型(gereral expression for linear and nonlinear auto-regressive mod-el,简称GNAR模型)的机械系统状态识别与故障诊断方法。利用采集系统工作过程中的特征信号建立GNAR模型;用主成分分析策略生成模型特征量,对训练样本的特征量进行识别和分类,得到各种参考模式;将几何距离判别函数作为状态分类的原则,根据待判系统特征量与各类参考模式的Euclide距离进行状态识别和故障判别。对车床颤振试验数据及高速离心空气压缩机故障数据的分析表明,该方法快捷、高效,诊断成功率较好,具有良好的工程应用前景。展开更多
This paper mainly deals with the Bayesian statistical inference theory on the VAR(p) forecasting model based on the parameters’ Minnesota conjugate prior distribution,including the prior distribution’s structure, th...This paper mainly deals with the Bayesian statistical inference theory on the VAR(p) forecasting model based on the parameters’ Minnesota conjugate prior distribution,including the prior distribution’s structure, the parameters’ posterior distribution, and compares the forecasting accuracy of AR,VAR and BVAR model.展开更多
文摘提出一种基于非线性自回归时间序列模型(gereral expression for linear and nonlinear auto-regressive mod-el,简称GNAR模型)的机械系统状态识别与故障诊断方法。利用采集系统工作过程中的特征信号建立GNAR模型;用主成分分析策略生成模型特征量,对训练样本的特征量进行识别和分类,得到各种参考模式;将几何距离判别函数作为状态分类的原则,根据待判系统特征量与各类参考模式的Euclide距离进行状态识别和故障判别。对车床颤振试验数据及高速离心空气压缩机故障数据的分析表明,该方法快捷、高效,诊断成功率较好,具有良好的工程应用前景。
文摘This paper mainly deals with the Bayesian statistical inference theory on the VAR(p) forecasting model based on the parameters’ Minnesota conjugate prior distribution,including the prior distribution’s structure, the parameters’ posterior distribution, and compares the forecasting accuracy of AR,VAR and BVAR model.
