计算机技术与发展2025，Vol.35Issue(5)：76-81,6.DOI:10.20165/j.cnki.ISSN1673-629X.2024.0395
基于DCNN-Informer的航空发动机寿命预测方法
Approach for Aircraft Engine Life Prediction Based on DCNN-Informer
摘要
Abstract
Prognostics and Health Management(PHM)plays a crucial role in industrial engineering,where predicting the remaining useful life(RUL)is essential for formulating maintenance strategies and reducing industrial losses,particularly for aircraft engines.Due to the increasing complexity of degradation characteristics in aircraft engines,leading to low accuracy in RUL prediction,we utilize conv-olutional neural networks(CNN)to extract high-dimensional spatial features of time series data,and globally model them in combination with Informer's self-attention mechanism,so as to fully extract time dimension information.Moreover,to further enhance the model's accuracy and generalization ability,a secondary training framework for engines is designed.This framework groups the dataset by engine and feeds each group's data into the CNN-Informer model for secondary training,resulting in personalized models for each engine.Finally,a second-order exponential smoothing filter algorithm is used to filter the model's prediction results,improving pre-diction stability and accuracy.Experimental results demonstrate that the proposed prediction model has significant advantages in RUL prediction compared to existing models,with superior predictive performance.关键词
深度学习/剩余寿命预测/Informer/卷积神经网络/航空发动机Key words
deep learning/remaining useful life prediction/Informer/convolutional neural networks/aircraft engines分类
信息技术与安全科学引用本文复制引用
廖雪超,陈海力,钟实..基于DCNN-Informer的航空发动机寿命预测方法[J].计算机技术与发展,2025,35(5):76-81,6.基金项目
国家自然科学基金项目(62273264) （62273264）
