摘要
提出了一种滚动轴承故障诊断的新方法。首次将自适应最稀疏时频分析(ASTFA)方法应用于振动信号的降噪,并针对KVPMCD方法只选择一种最佳相关模型而忽略其他几种相关模型对预测精度贡献的缺陷,提出了一种改进的KVPMCD模式识别算法——人工鱼群算法优化融合Kriging模型的基于变量预测模型的模式识别(AKVPMCD)算法,即采用收敛速度快、鲁棒性强、具有全局寻优能力的人工鱼群智能算法(AFSIA)优化融合多种Kriging相关模型来提高模型预测精度。在此基础上,提出了一种基于ASTFA降噪和AKVPMCD算法的滚动轴承故障诊断方法。实验结果表明,该方法可以有效提高分类识别的精度。
A new rolling bearing fault diagnosis method was proposed. ASTFA method was ap- plied to the vibration signal de-noising for the first time.Aiming at the defects of KVPMCD(Kriging model variable predictive model based class discriminate) method that was chosen one of the best re- lated model only and the other correlation models' contribution to the prediction accuracy was ig- nored,an improved KVPMCD pattern recognition algorithm AKVPMCD was proposed, the AFSIA (artificial fish swarm intelligence algorithm) which had high convergence speed, strong robustness and global optimization ability was used to optimize a variety of Kriging models, so as to improve the prediction precision. On the basis of above, a new fault diagnosis method of rolling bearings was pro- posed based on ASTFA de-noising and AKVPMCD. The experimental results prove that this method can improve the precision of classification recognition effectively.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2015年第21期2934-2940,共7页
China Mechanical Engineering
基金
国家自然科学基金资助项目(51175158
51375152)
湖南省自然科学基金资助项目(11JJ2026)
关键词
自适应最稀疏时频分析降噪
AKVPMCD
滚动轴承
故障诊断
adaptive and sparsest time-frequency analysis (ASTFA) de-noising
artificial fish swarm algorithm optimizing fusion Kriging model variable predictive model based class discriminate (AKVPMCD)
rolling bearing
fault diagnosis