摘要
提出了一种门限自回归(AR)模型的盲辨识算法,并与常用方法进行比较分析。该算法的特点在保证辨识精度上可大大提高其运行速度,而且阶次越高,该算法的优势越明显。将该方法与隐Markov模型结合,以门限自回归模型各区间的AR子模型系数作为特征向量,以隐Markov模型作为分类器,应用到旋转机械升降速过程的故障诊断中。实验结果表明,这种方法有很好的实用性。
A blind identification method is developed for the threshold auto-regressive (AR) model. The method has good identification accuracy and rapid convergence, especially for higher order systems. The method was then combined with the Hidden Markov model to determine the AR coefficients for each interval used for feature extraction, with the Hidden Markov model as a classifier. Tests show that the system can be used for fault diagnosis during the process of startup and slowing down of rotating machinery.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2005年第8期1036-1039,共4页
Journal of Tsinghua University(Science and Technology)
基金
中国博士后基金项目(20040350061)
国家自然科学基金资助项目(50105007)
教育部"跨世纪优秀人才培养计划"基金资助
关键词
通讯理论
门限自回归模型
盲辨识
故障诊断
隐MARKOV模型
communication theory
threshold auto-regressive model(TAR)
blind identification
fault diagnosis
hidden Markov mode (HMM)