전이학습 기반 CNN-LSTM을 통한 항공기 터보팬 잔여 유효 수명 예측
Prediction Remaining Useful Life of Aircraft Turbofans Using Transfer Learning Based CNN-LSTM
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
항공기 엔진의 잔여 유효 수명(RUL) 예측은 고장 예지 및 건전성 관리(PHM)를 좌우하는 중요한 역할이며 관련 연구가 활발히 진행되고 있다. 하지만 실제 고장 데이터를 확보하는 데 어려움이 있으며 결함 데이터의 희소성으로 인해 예측 모델 구축에 한계가 존재한다. 기존의 데이터 증대방법은 제한된 결함 데이터를 보완하기 위해 사용되지만, 휴리스틱한 접근법으로 인해 데이터 특성이 일반화되는 문제가 발생하며 예측 정확도가저하될 수 있다. 이러한 문제를 해결하기 위해, 본 연구에서는 전이학습(Transfer Learning)에 기반한 CNN-LSTM 혼합모델을 제안한다. CNN-LSTM혼합모델은 CNN의 특징 추출 기능과 LSTM의 장기 시계열 학습 능력을 결합하여, 항공기 엔진 센서 데이터의 복잡한 동적 특성과 시계열적 변화를동시에 반영한다. 본 모델로 서로 다른 소스 도메인(Source Domain)에서 학습된 데이터를 전이학습을 통해 타겟 도메인(Target Domain) 데이터에적용함으로써, 예측 기법의 성능을 향상시킨다. 미국항공우주국(NASA)에서 개발한 C-MAPSS 항공기 엔진 시뮬레이터 데이터 세트에 적용한 결과,전이학습을 통한 사전 학습된 모델의 활용이 기본 혼합모델보다 예측 정확도 향상에 기여함을 확인하였다. 또한, 제안된 모델이 PHM 분야에서주요한 여러 예측 모델들 대비 우수한 예측 성능을 보임을 확인하였다.
Abstract
The objective of research in the field of prognostics and health management is to predict the Remaining Useful Life of aircraft engines,a critical component of analysis within this domain. Nevertheless, there are difficulties in acquiring dependable failure information, andthe limited availability of defect data hinders the development of predictive models. Current data augmentation techniques are utilizedto enhance the insufficient defect data; however, the heuristic approaches might oversimplify the data characteristics, ultimately decreasingpredictive accuracy. This study suggests a hybrid model that combines Transfer Learning, specifically integrating Convolutional NeuralNetworks (CNN) and Long Short-Term Memory (LSTM). The hybrid CNN-LSTM model integrates the CNN’s feature extraction capabilitieswith the LSTM’s long-term time series learning capacity, facilitating the representation of intricate dynamic characteristics and temporalfluctuations in aircraft engine sensor data. The performance of predictive techniques is enhanced by applying data learned from varioussource domains to target domain data through transfer learning. The results obtained by applying this model to the C-MAPSS aircraftengine simulator dataset developed by the National Aeronautics and Space Administration (NASA) corroborate the idea that employinga pre-trained model through transfer learning improves predictive accuracy in comparison to the standard mixed model. Furthermore,the proposed model demonstrates improved predictive abilities when compared to various leading predictive models in the PHM field.
참고문헌 (35)
- O. Asif, "A Deep Learning Model for Remaining Useful Life Prediction of Aircraft Turbofan Engine on C-MAPSS Dataset", IEEE Access, 10, pp.95425~59440 (2022)[DOI]
- H. Gharib, "A Review of Prognostic and Health Management(PHM)Methods and Limitations for Marine Diesel Engines : New Research Directions", Machines, 11(7), pp.695~ (2023)
- J. W. Baek, "Prognostics and Health Management using Artificial Intelligence", Journal of Reliability Society of Korea, 23(3), pp.283~296 (2023)
- X. Li, "Remaining Useful Life Estimation in Prognostics Using Deep Convolution Neural Networks", Reliability Engineering &System Safety, 172, pp.1~11 (2018)
- C. Shorten, "A Survey on Image Data Augmentation for Deep Learning", Journal of Big Data, 6(1), pp.60~ (2019)
- E. Eldele, "Label-Efficient Time Series Representation Learning : A Review", IEEE Transactions on Artificial Intelligence, 5(1), pp.1~16 (2024)
- A. Zhang, "Transfer Learning with Deep Recurrent Neural Networks for Remaining Useful Life Estimation", Applied Sciences, 8(12), pp.2416~ (2018)
- M. Gribbestad, "Transfer Learning for Prognostics and Health Management(PHM)of Marine Air Compressors", Journal of Marine Science and Engineering, 9(1), pp.47~ (2021)
- Y. Duan, "Cross-Domain Remaining Useful Life Prediction Based on Adversarial Training", Machines, 10(6), pp.438~ (2022)
- Y. Ding, "Transfer Learning for Remaining Useful Life Prediction Across Operating Conditions Based on Multisource Domain Adaptation", IEEE, 2022, pp.27(5), 4143~4152
- Y. Hu, "A Novel Remaining Useful Life Probability Prediction Approach for Aero-Engine with Improved Bayesian Uncertainty Estimation Based on Degradation Data", Applied Sciences, 13(16), pp.9194~ (2023)
- T. Tang, "Transfer Learning for Bearing Fault Diagnosis Based on Graph Neural Network with Dilated KNN and Adversarial Discriminative Domain Adaptation", Measurement Science and Technology, 35(6), pp.065106~ (2024)
- Y. Hsueh, "Condition Monitor System for Rotation Machine by CNN with Recurrence Plot", Energies, 12(17), pp.3221~ (2019)
- Y. J. Seo, "An Input Transformation with MFCCs and CNN Learning Based Robust Bearing Fault Diagnosis Method for Various Working Conditions", KIPS Transactions on Software and Data Engineering, 11(4), pp.179~188 (2022)
- J. H. Baek, "Gear Fault Diagnosis Based on Residual Patterns of Current and Vibration Data by Collaborative Robot’s Motions Using LSTM", The Transactions of the Korea Information Processing Society, 12(10), pp.445~454 (2023)
- H. Lu, "A Multi-Source Transfer Learning Model Based on LSTM and Domain Adaptation for Building Energy Prediction", International Journal of Electrical Power & Energy Systems, 149, pp.109024~ (2023)
- Z. Liu, "A Study of a Domain-Adaptive LSTM-DNN-Based Method for Remaining Useful Life Prediction of Planetary Gearbox", Processes, 11(7), pp.2002~ (2023)
- S. Seo, "Supervised Domain Adaptation for Remaining Useful Life Prediction Based on AdaBoost With Long Short-Term Memory", IEEE Access, 12, pp.96757~96768 (2024)
- P. Khumprom, "A Hybrid Evolutionary CNN-LSTM Model for Prognostics of C-MAPSS Aircraft Dataset", 2023 Annual Reliability and Maintainability Symposium(RAMS), pp, pp.1~8 (2023)
- N. Borst, "Introducing CNN-LSTM Network Adaptations to Improve Remaining Useful Life Prediction of Complex Systems", The Aeronautical Journal, 127(1318), pp.2143~2153 (2023)
- H. Mo, "Multi-Head CNN-LSTM with Prediction Error Analysis for Remaining Useful Life Prediction", Proceedings of the 27th Conference of Open Innovations Association (FRUCT), pp.164~171 (2020)
- S. Deng, "Prediction of Remaining Useful Life of Aero-Engines Based on CNN-LSTM-Attention", International Journal of Computational Intelligence Systems, 17(1), pp.232~ (2024)
- User’s Guide for the Commercial Modular Aero-Propulsion System Simulation(C-MAPSS), "D. K. Frederick", NASA Glenn Research Center, ~ (2007)
- L. Liu, "Entropy-based sensor selection for condition monitoring and prognostics of aircraft engine", Microelectronics Reliability, 55(9-10), pp.2092~2096 (2015)
- T. Berghout, "A Systematic Guide for Predicting Remaining Useful Life with Machine Learning", Electronics, 11(7), pp.1125~ (2022)
- M. T. Furqon, "Mixup Domain Adaptations for Dynamic Remaining Useful Life Predictions", Knowledge-Based Systems, 295, pp.111783~ (2024)
- C. Zhao, "A Double-Channel Hybrid Deep Neural Network Based on CNN and BiLSTM for Remaining Useful Life Prediction", Sensors, 20(24), pp.7109~ (2020)
- F. Li, "A Light Gradient Boosting Machine for Remaining Useful Life Estimation of Aircraft Engines", Proceedings of the 21st International Conference on Intelligent Transportation Systems (ITSC), pp.3562~3567 (2018)
- A. F. Azyus, "Prediction of Remaining Useful Life Using the CNN-GRU Network : A Study on Maintenance Management", Software Impacts, 17, pp.100535~ (2023)
- G. S. Babu, "Deep Convolutional Neural Network-Based Regression Approach for Estimation of Remaining Useful Life", Proceedings of the 21st International Conference on Database Systems for Advanced Applications (DASFAA 2016), 9642, pp.214~228 (2016), Lecture Notes in Computer Science
- X. Li, "Remaining Useful Life Prediction with Partial Sensor Malfunctions Using Deep Adversarial Networks", IEEE, 2023, pp.10(1), 121~134
- C. Louen, "A New Framework for Remaining Useful Life Estimation Using Support Vector Machine Classifier", Proceedings of IEEE Conference on Control and Fault-Tolerant Systems, pp.228~233 (2013)
- P. R. de O. da Costa, "Remaining Useful Lifetime Prediction via Deep Domain Adaptation", Reliability Engineering &System Safety, 195, pp.106682~ (2020)
- M. Ragab, "Contrastive Adversarial Domain Adaptation for Machine Remaining Useful Life Prediction", IEEE Transactions on Industrial Informatics, 17(8), pp.5239~5249 (2021)