基于Stacking多模型融合的航空发动机剩余使用寿命预测方法

白伟伟; 杨治宇; 王晓敏; 柴进

doi:10.7643/ issn.1672-9242.2026.05.007

PDF(1315 KB)

装备环境工程 ›› 2026, Vol. 23 ›› Issue (5) : 53-63. DOI: 10.7643/ issn.1672-9242.2026.05.007

航空航天装备

基于Stacking多模型融合的航空发动机剩余使用寿命预测方法

白伟伟¹, 杨治宇², 王晓敏¹, 柴进¹

作者信息 +

Remaining Useful Life Prediction Method of Aero Engines Based on Stacking Multi-model Fusion

BAI Weiwei¹, YANG Zhiyu², WANG Xiaomin¹, CHAI Jin¹

Author information +

文章历史 +

摘要

目的解决航空发动机剩余使用寿命（RUL）预测中单一模型泛化能力不足的问题。方法提出一种基于Stacking多模型融合的航空发动机剩余使用寿命预测方法。首先利用随机森林算法和最大互信息系数对航空发动机剩余使用寿命预测的关键特征进行筛选,在此基础上,采用滑动窗口方法对数据集中发动机运行数据进行样本重构,以提取的关键特征作为模型输入,以RUL值作为输出,构建融合多个单一机器学习算法的Stacking预测模型。结果相对于单一机器学习算法构建的模型,所提出的Stacking模型表现较好,相较于最优基学习器,Stacking模型RMSE和Score值分别减少19.16%和38.52%。消融实验证明特征筛选可以提升Stacking模型准确率。结论经过特征筛选的Stacking多模型融合方法可以有效提升RUL值的预测精度。

Abstract

The work aims to address the issue of insufficient generalization ability of a single model in predicting the remaining useful life (RUL) of aero engines. A remaining useful life prediction method of aero engines based on Stacking multi-model fusion was proposed. Firstly, the key features for predicting the RUL of aero engines were selected with the random forest algorithm and the maximal information coefficient. Based on this, the sliding window method was used to reconstruct engine operating data samples in the dataset. With the extracted key features as model input and RUL as output, a Stacking prediction model that integrated multiple individual machine learning algorithms was built. Compared with the model constructed by a single machine learning algorithm, the proposed Stacking model performed better. Compared with the optimal base learner, the RMSE and Score values of the Stacking model were reduced by 19.16% and 38.52%, respectively. The ablation experiment proved that feature screening could improve the accuracy of the Stacking model. It is proved that the Stacking multi-model fusion method with feature screening can effectively improve the prediction accuracy of RUL.

导出引用

白伟伟, 杨治宇, 王晓敏, 柴进. 基于Stacking多模型融合的航空发动机剩余使用寿命预测方法[J]. 装备环境工程. 2026, 23(5): 53-63 https://doi.org/10.7643/ issn.1672-9242.2026.05.007

BAI Weiwei, YANG Zhiyu, WANG Xiaomin, CHAI Jin. Remaining Useful Life Prediction Method of Aero Engines Based on Stacking Multi-model Fusion[J]. Equipment Environmental Engineering. 2026, 23(5): 53-63 https://doi.org/10.7643/ issn.1672-9242.2026.05.007

中图分类号： V235.13

参考文献

[1] 杨阳, 李中生, 王艺, 等. 利用性能参数回归分析的航空发动机MTBF预计[J]. 航空发动机, 2023, 49(1): 34-40.
YANG Y, LI Z S, WANG Y, et al.Prediction of Aeroengine MTBF Using Regression Analysis of Performance Parameters[J]. Aeroengine, 2023, 49(1): 34-40.
[2] 陈俊英, 席月芸, 徐琳, 等. 基于MLP集成随机子空间决策树的航空发动机剩余使用寿命预测[J]. 航空发动机, 2024, 50(6): 81-87.
CHEN J Y, XI Y Y, XU L, et al.Remaining Useful Life Prediction of Aeroengines Based on MLP Integrated Random Subspace Decision Trees[J]. Aeroengine, 2024, 50(6): 81-87.
[3] 齐阳, 祝华远, 吴士博. 基于VMD-LSTM的航空发动机剩余寿命预测方法[J]. 装备环境工程, 2025, 22(5): 92-102.
QI Y, ZHU H Y, WU S B.Residual Useful Life Prediction Method of Aero Engine Based on VMD-LSTM[J]. Equipment Environmental Engineering, 2025, 22(5): 92-102.
[4] 秦子轩, 张晓东, 白广芝, 等. 基于多尺度特征融合的航空发动机剩余寿命预测[J]. 航空发动机, 2024, 50(4): 114-120.
QIN Z X, ZHANG X D, BAI G Z, et al.Aeroengine Remaining Useful Life Prediction Based on Multi-Scale Feature Fusion[J]. Aeroengine, 2024, 50(4): 114-120.
[5] 孙秀慧, 李娟, 戴洪德. 基于Copula熵传感器选择的发动机相似性寿命预测方法[J]. 航空发动机, 2024, 50(5): 113-121.
SUN X H, LI J, DAI H D.Engine Similarity Life Prediction Based on Copula Entropy for Sensor Selection[J]. Aeroengine, 2024, 50(5): 113-121.
[6] 周先烨, 陈里里, 王彪, 等. 基于特征重叠与集成学习的涡扇发动机剩余寿命预测[J]. 自动化与仪表, 2023, 38(8): 51-55.
ZHOU X Y, CHEN L L, WANG B, et al.Residual Lifetime Prediction of Turbofan Engine Based on Features Overlap and Ensemble Learning[J]. Automation & Instrumentation, 2023, 38(8): 51-55.
[7] 张一鸣, 刘晓锋, 崔宝, 等. 基于集成学习的涡扇发动机剩余寿命预测模型[J]. 热能动力工程, 2022, 37(4): 64-70.
ZHANG Y M, LIU X F, CUI B, et al.Turbofan Engine Remaining Useful Life Prediction Model Based on Ensemble Learning[J]. Journal of Engineering for Thermal Energy and Power, 2022, 37(4): 64-70.
[8] 张一鸣, 刘晓锋. 基于多特征提取和Stacking集成学习的航空发动机余寿预测[C]// 第六届空天动力联合会议暨中国航天第三专业信息网第四十二届技术交流会暨2021航空发动机技术发展高层论坛. 北京: 出版者不详, 2022.
ZHANG Y M, LIU X F.An Aero-engine RUL Prediction Model Based on Feature Selection and Stacking Ensemble Learning Model[C]// Proceedings of the 6th Aerospace Power Joint Conference and the 42nd Technical Exchange Conference of China Aerospace Third Professional Information Network and the 2021 Advanced Forum on the Development of Aeroengine Technology. Beijing: [s. n.], 2022.
[9] 张豹, 刘琼, 吴细宝, 等. 基于集成学习的涡扇发动机剩余寿命预测[J]. 中国测试, 2022, 48(7): 47-52.
ZHANG B, LIU Q, WU X B, et al.Remaining Useful Lifetime Prediction of Turbofan Engine Based on Ensemble Learning[J]. China Measurement & Test, 2022, 48(7): 47-52.
[10] 吴兴华, 包俊, 王洪亮. 基于DBN-SO-BiGRU融合模型的航空发动机RUL预测[J/OL]. 控制工程, 2024: 1-10. (2024-05-14)[2024-09-05]. https://doi.org/10.14107/j.cnki.kzgc.20240328.
WU X H, BAO J, WANG H, et al. Aero-engine RUL Prediction Based on DBN-SO-BiGRU Fusion Model[J/OL]. Control Engineering of China, 2024: 1-10. (2024-05-14)[2024-09-05]. https://doi.org/10.14107/j.cnki.kzgc.20240328.
[11] 胡艳艳, 白雅婷. 基于贝叶斯图注意力Transformer的航空发动机剩余使用寿命概率预测[J]. 工程科学学报, 2025, 47(2): 374-388.
HU Y Y, BAI Y T.Probabilistic Prediction of Aero Engine Remaining Useful Life Based on Bayesian Graph Attention Transformer[J]. Chinese Journal of Engineering, 2025, 47(2): 374-388.
[12] 段佳俊, 陆中, 王捷. 基于多尺度CNN和Transformer的航空发动机寿命预测方法[J]. 航空动力学报, 2026, 41(3): 311-321.
DUAN J J, LU Z, WANG J.Aero-Engine Life Prediction Method Based on Multi-Scale CNN and Transformer[J]. Journal of Aerospace Power, 2026, 41(3): 311-321.
[13] 杨森, 吴素萍. 基于循环神经网络和集成学习的气温预报误差订正方法与应用[J]. 农业工程学报, 2025, 41(20): 158-167.
YANG S, WU S P.Temperature Forecast Error Correction Method and Application Based on Recurrent Neural Network and Ensemble Learning[J]. Transactions of the Chinese Society of Agricultural Engineering, 2025, 41(20): 158-167.
[14] 敖渝焱. 基于机器学习的PNP方程迭代求解的参数选取研究[D]. 桂林: 桂林电子科技大学, 2025.
AO Y Y.The Study on Parameter Selection for Iterative Solution of the PNP Equations Based on Machine Learning[D]. Guilin: Guilin University of Electronic Technology, 2025.
[15] 张源远, 刘雪平, 叶春林, 等. 基于VMD-SMA-SVR的改扩建道路沉降预测模型研究[J]. 科技和产业, 2025, 25(23): 117-123.
ZHANG Y Y, LIU X P, YE C L, et al.Research on Settlement Prediction Model for Reconstruction and Expansion Roads Based on VMD-SMA-SVR[J]. Science Technology and Industry, 2025, 25(23): 117-123.
[16] 王小川, 史峰, 郁磊, 等. MATLAB神经网络43个案例分析[M]. 北京: 北京航空航天大学出版社, 2013: 1-2.
WANG X C, SHI F, YU L, et al.Analysis of 43 Cases of MATLAB Neural Network[M]. Beijing: Beihang University Press, 2013: 1-2.
[17] LUO S C, WANG B S, GAO Q Z, et al.Stacking Integration Algorithm Based on CNN-BiLSTM-Attention with XGBoost for Short-Term Electricity Load Forecasting[J]. Energy Reports, 2024, 12: 2676-2689.
[18] MOON J, JUNG S, REW J, et al.Combination of Short-Term Load Forecasting Models Based on a Stacking Ensemble Approach[J]. Energy and Buildings, 2020, 216: 109921.
[19] WOLPERT D H.Stacked Generalization[J]. Neural Networks, 1992, 5(2): 241-259.
[20] BREIMAN L.Stacked Regressions[J]. Machine Learning, 1996, 24(1): 49-64.
[21] 郑颖颖, 李鑫, 陈延旭, 等. 基于Stacking多模型融合的极端天气短期风电功率预测方法[J]. 高电压技术, 2024, 50(9): 3871-3882.
ZHENG Y Y, LI X, CHEN Y X, et al.Short-Term Wind Power Forecasting Method in Extreme Weather Based on Stacking Multi-Model Fusion[J]. High Voltage Engineering, 2024, 50(9): 3871-3882.
[22] 马良玉, 耿妍竹, 梁书源, 等. 基于Stacking多模型融合的风电机组齿轮箱油池温度异常预警[J]. 中国电机工程学报, 2023, 43(S1): 242-251.
MA L Y, GENG Y Z, LIANG S Y, et al.Anomaly Warning of Wind Turbine Gearbox Oil Pool Temperature Based on Stacking Fusion of Multiple Models[J]. Proceedings of the CSEE, 2023, 43(S1): 242-251.
[23] 吴俊华, 王粮局, 徐际童, 等. 基于Stacking多模型融合的颗粒饲料质量预测方法[J]. 农业工程学报, 2025, 41(15): 318-326.
WU J H, WANG L J, XU J T, et al.Pellet Feed Quality Prediction Method Based on Stacking Multi-Model Fusion[J]. Transactions of the Chinese Society of Agricultural Engineering, 2025, 41(15): 318-326.
[24] 沈赋, 张微, 徐潇源, 等. 基于随机森林和最大互信息系数关键特征选择的配电网拓扑辨识研究[J]. 电力系统保护与控制, 2024, 52(17): 1-11.
SHEN F, ZHANG W, XU X Y, et al.Topological Identification of Distribution Networks Based on Key Feature Selection Using RF and MIC[J]. Power System Protection and Control, 2024, 52(17): 1-11.
[25] 丁子尧. 涡扇发动机性能衰退状态与剩余寿命预测方法[D]. 大连: 大连理工大学, 2023.
DING Z Y.Performance Decline State and Residual Life Prediction Method for Turbofan Engines[D]. Dalian: Dalian University of Technology, 2023.
[26] SAXENA A, GOEBEL K, SIMON D, et al.Damage Propagation Modeling for Aircraft Engine Run-to-Failure Simulation[C]//2008 International Conference on Prognostics and Health Management. Denver: IEEE, 2008.
[27] 张诺飞. 基于脉冲神经网络及深度学习的航空发动机剩余使用寿命预测研究[D]. 石家庄: 河北师范大学, 2025: 37-43.
ZHANG N F.Research on Prediction of Aircraft Engines' Remaining Useful Life Based on Spiking Neural Networks and Deep Learning[D]. Shijiazhuang: Hebei Normal University, 2025: 37-43.
[28] 王昊. 航空发动机剩余使用寿命预测研究[D]. 大连: 大连理工大学, 2022.
WANG H.Research on Remaining Useful Life Prediction of Aeroengine[D]. Dalian: Dalian University of Technology, 2022.
[29] WANG T Y, YU J B, SIEGEL D, et al.A Similarity-Based Prognostics Approach for Remaining Useful Life Estimation of Engineered Systems[C]//2008 International Conference on Prognostics and Health Management. Denver: IEEE, 2008.
[30] 王昆, 郭迎清, 赵万里, 等. 基于SSAE和相似性匹配的航空发动机剩余寿命预测[J]. 北京航空航天大学学报, 2023, 49(10): 2817-2825.
WANG K, GUO Y Q, ZHAO W L, et al.Remaining Useful Life Prediction of Aeroengine Based on SSAE and Similarity Matching[J]. Journal of Beijing University of Aeronautics and Astronautics, 2023, 49(10): 2817-2825.
[31] 李杰, 贾渊杰, 张志新, 等. 基于融合神经网络的航空发动机剩余寿命预测[J]. 推进技术, 2021, 42(8): 1725-1734.
LI J, JIA Y J, ZHANG Z X, et al.Remaining Useful Life Prediction of Aeroengine Based on Fusion Neural Network[J]. Journal of Propulsion Technology, 2021, 42(8): 1725-1734.
[32] GONG R, LI J X, WANG C L.Remaining Useful Life Prediction Based on Multisensor Fusion and Attention TCN-BiGRU Model[J]. IEEE Sensors Journal, 2022, 22(21): 21101-21110.
[33] AYDEMIR G, ACAR B.Anomaly Monitoring Improves Remaining Useful Life Estimation of Industrial Machinery[J]. Journal of Manufacturing Systems, 2020, 56: 463-469.