航空学报 > 2023, Vol. 44 Issue (11): 227660-227660   doi: 10.7527/S1000-6893.2022.27660

基于Box-Cox变换与随机系数回归模型的非线性退化设备剩余寿命预测方法

杨保奎, 张建勋, 李慧琴, 司小胜()   

  1. 火箭军工程大学 智剑实验室,西安 710025
  • 收稿日期:2022-06-22 修回日期:2022-07-14 接受日期:2022-07-31 出版日期:2023-06-15 发布日期:2022-08-08
  • 通讯作者: 司小胜 E-mail:sxs09@mails.stinghua.edu.cn;sxs09@mails.tsinghua.edu.cn
  • 基金资助:
    国家自然科学基金(61922089)

Remaining useful life prediction method for nonlinear degrading equipment based on Box-Cox transformation and random coefficient regression model

Baokui YANG, Jianxun ZHANG, Huiqin LI, Xiaosheng SI()   

  1. Zhijian Laboratory,Rocket Force University of Engineering,Xi’an 710025,China
  • Received:2022-06-22 Revised:2022-07-14 Accepted:2022-07-31 Online:2023-06-15 Published:2022-08-08
  • Contact: Xiaosheng SI E-mail:sxs09@mails.stinghua.edu.cn;sxs09@mails.tsinghua.edu.cn
  • Supported by:
    National Natural Science Foundation of China(61922089)

摘要:

准确预测退化设备的剩余寿命可以为设备维护管理提供重要信息支撑,进而避免设备运行中发生计划外失效、减少设备运行维护成本。针对工程实际设备广泛存在的非线性退化现象,提出了基于Box-Cox变换与随机系数回归模型的非线性退化设备剩余寿命预测方法。首先,采用Box-Cox变换对非线性退化数据进行线性化处理,在此基础上通过随机系数回归模型构建退化模型,并运用Bayesian理论与蒙特卡洛-期望最大化算法在线更新模型参数;然后,基于随机系数回归模型的特性,推导出剩余寿命的分布函数以及其点估计值;最后,通过数值仿真和锂电池实际退化数据验证所提方法的有效性。

关键词: 剩余寿命, Box-Cox变换, 随机系数回归模型, 非线性退化数据, 蒙特卡洛-期望最大化算法

Abstract:

Accurate prediction of Remaining Useful Life (RUL) of degraded equipment can provide important information support for equipment maintenance management, thereby avoiding unplanned failure and reducing the operating cost of equipment. Aiming at the nonlinear degradation phenomenon widely existing in practical engineering, this paper proposes a RUL prediction method of nonlinear equipment based on the Box-Cox transformation and random coefficient regression model. The Box-Cox transformation is used to linearize the nonlinear degradation data, the degradation model is then constructed through the random coefficient regression model based on the transformed degradation data, and the model parameters are updated online by the Bayesian theory and Monte Carlo expected maximization algorithm. Based on the characteristics of the random coefficient regression model, the distribution function of RUL and its point estimation value are derived. Finally, the effectiveness of the proposed method is verified by numerical simulation and actual degradation data of a lithium battery.

Key words: remaining useful life (RUL), Box-Cox transformation, random coefficient regression model, nonlinear degradation data, Monte Carlo expected maximization algorithm

中图分类号: