CAJ | 학술논문

数据预处理是基于相似性的寿命预测方法的首要步骤，重点研究了变工况下数据预处理的方法。首先研究了其必要性，随后从线性模型和非线性模型的角度研究了多工况健康评估方法。从数据标准化、基于经验信噪比的特征加权和基于主成分分析的特征降维研究了多工况数据标准化方法。研究证明，多工况数据标准化方法更适于相似性寿命预测中的数据预处理。
수거예처리시기우상사성적수명예측방법적수요보취，중점연구료변공황하수거예처리적방법。수선연구료기필요성，수후종선성모형화비선성모형적각도연구료다공황건강평고방법。종수거표준화、기우경험신조비적특정가권화기우주성분분석적특정강유연구료다공황수거표준화방법。연구증명，다공황수거표준화방법경괄우상사성수명예측중적수거예처리。
Data preprocessing is the first step of life prediction method based on similarity . T he data preprocessing method under variable condition is researched in this paper .The necessity is researched firstly ,then the muli‐regime health assement is studied in the perspective of linear model and nonlinear model .The muti‐regime data normalization is studied in the perspective of data normalization ,variable weighting based on experience signal‐to‐noise ratio and variable reduction based on PCA .It shows that the muti‐regime data normalization is more suitable for muti‐regime data in RUL prediction based on similarity .