• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2021, Vol. 57 ›› Issue (22): 106-113.doi: 10.3901/JME.2021.22.106

• 仪器科学与技术 • 上一篇    下一篇

扫码分享

数字孪生驱动的航空发动机涡轮盘剩余寿命预测

付洋1, 曹宏瑞1,2, 郜伟强3, 高文辉3   

  1. 1. 西安交通大学航空发动机研究所 西安 710049;
    2. 西安交通大学机械制造系统工程国家重点实验室 西安 710049;
    3. 中国燃气涡轮研究院强度技术研究室 成都 610500
  • 收稿日期:2020-11-11 修回日期:2021-06-12 出版日期:2021-11-20 发布日期:2022-02-28
  • 通讯作者: 曹宏瑞(通信作者),男,1982年出生,博士,教授,博士研究生导师。主要研究方向为机械系统建模与动态信号处理、机械设备健康监测与智能维护、数字孪生。E-mail:chr@mail.xjtu.edu.cn
  • 作者简介:付洋,男,1992年出生,博士研究生。主要研究方向为机械设备故障诊断与寿命预测。E-mail:yangfu@stu.xjtu.edu.cn
  • 基金资助:
    国家自然科学基金优秀青年科学基金(51922084)和航空动力基金(6141B09050394)资助项目。

Digital Twin Driven Remaining Useful Life Prediction for Aero-engine Turbine Discs

FU Yang1, CAO Hongrui1,2, GAO Weiqiang3, GAO Wenhui3   

  1. 1. Institute of Aero-engine, Xi'an Jiaotong University, Xi'an 710049;
    2. State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University, Xi'an 710049;
    3. The Strength Research Office, China Gas Turbine Establishment, Chengdu 610500
  • Received:2020-11-11 Revised:2021-06-12 Online:2021-11-20 Published:2022-02-28

摘要: 为解决航空发动机涡轮盘剩余寿命在线预测难题,提出一种数字孪生驱动的涡轮盘剩余寿命预测方法。在建立数字孪生模型的过程中,首先,分析涡轮盘疲劳裂纹损伤机理,构建性能退化指标,建立涡轮盘性能退化过程的共性表征模型;其次,分析多种不确定性因素,采用状态空间模型建立涡轮盘性能退化过程的个性表征模型;然后,通过动态贝叶斯网络描述状态空间模型随时间的演化规律,建立涡轮盘性能退化过程的动态演化模型;最后,采用粒子滤波算法实现涡轮盘退化状态追踪和剩余寿命预测,从而完成涡轮盘性能退化数字孪生模型的建立。融合涡轮盘实时传感数据,通过贝叶斯推理实现对该数字孪生模型的动态更新。通过某型涡轮盘试验数据对该方法进行验证,结果表明该数字孪生模型能够较好地解决涡轮盘剩余寿命在线预测问题。

关键词: 涡轮盘, 性能退化, 数字孪生, 剩余寿命预测

Abstract: A digital twin (DT) based remaining useful life (RUL) prediction method is proposed for on-line RUL prediction of aero-engine turbine disc. In the proposed DT model, a common representation model is first developed to depict the performance degradation process of the turbine disc based on the fatigue damage mechanism. Then, an individual representation model is established by using the state-space model with uncertainty analysis. Next, dynamic Bayesian network is employed to construct the dynamic evolution model, which depicts the dynamic performance degradation process of turbine disc. Finally, particle filter is used to make the DT model capable of tracking the performance degradation and predicting the RUL for turbine disc. Specially, the real-time sensor data is collected to update the DT model by Bayesian inference algorithm. The fatigue life test of turbine disc is carried out to validate the effectiveness of the proposed method. The results show that the DT model is capable to solve the on-line RUL prediction problem of turbine disc.

Key words: turbine disc, performance degradation, digital twin, remaining useful life prediction

中图分类号: