Hostname: page-component-5db58dd55d-8mwbx Total loading time: 0 Render date: 2026-05-25T14:22:40.323Z Has data issue: false hasContentIssue false
  • English
  • Français

An improved deep autoencoder framework for aviation piston engine unsupervised anomaly detection considering the real-life data characteristics

Published online by Cambridge University Press:  30 April 2026

Guo Li ,
Tongge Xu Open the ORCID record for Tongge Xu [Opens in a new window]  and
Shuiting Ding
Guo Li
Affiliation:
Beihang University, China
Tongge Xu*
Affiliation:
Beihang University, China
Shuiting Ding
Affiliation:
Civil Aviation University of China, China
*
Corresponding author: Tongge Xu; Email: xutongge114@outlook.com
Get access Rights & Permissions [Opens in a new window]

Abstract

Anomaly detection in piston engines plays a pivotal role in maintaining the continuous airworthiness and operational safety of general aviation aircraft. However, real-life data present several challenges, including diverse operating conditions, intermixed normal and abnormal samples and complex fault phenomena, all of which complicate model training and manual labeling tasks. To address the challenges, an improved deep autoencoder-based framework under multiple working conditions (IMDAEF) is proposed to enhance the accuracy of anomaly detection and the performance evaluation process. First, in the proposed method, a clustering-guided approach is introduced for working condition identification. Piston engine parameters are used for K-means clustering, and aircraft flight phase information is incorporated to fine-tune the results. Second, during the deep autoencoder training process, a sample selection mechanism and an equivalent threshold method are conducted, considering the presence of outliers in normal operation data. Finally, during the model evaluation stage, recognising the complexity of real-life fault phenomena that hinder manual labeling, an outlier region division and selection mechanism is proposed to re-annotate subtle fault occurrences in the fault datasets, thereby improving model evaluation accuracy. Results from four representative fault datasets confirm that the proposed framework maintains strong detection performance against complex faults. Moreover, the experiments demonstrate its ability to mitigate the negative impact of real-life data characteristics while meeting real-time performance requirements, offering valuable insights for future research.

Keywords

aviation piston enginedeep autoencoderreal-life datasafetyunsupervised anomaly detection

Information

Type
Research Article
Information
The Aeronautical Journal , First View , pp. 1 - 26
Check for updates
Copyright
© The Author(s), 2026. Published by Cambridge University Press on behalf of Royal Aeronautical Society

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

Chandola, V., Banerjee, A. and Kumar, V. Anomaly detection: a Survey, ACM Comput. Surv., 2009, 41, (3), pp 1–58.10.1145/1541880.1541882CrossRefGoogle Scholar
Khattak, W.R., Salman, A., Ghafoor, S. and Latif, S. Multi-modal LSTM network for anomaly prediction in piston engine aircraft, Heliyon, 2024, 10, (3), p e25120.10.1016/j.heliyon.2024.e25120CrossRefGoogle Scholar
Jaw, L.C. Recent advancements in aircraft engine health management (EHM) technologies and recommendations for the next step, Proceedings of ASME Turbo Expo 2005: Power for Land, Sea, and Air, Nevada, USA, 2005.10.1115/GT2005-68625CrossRefGoogle Scholar
Zhou, X., Lu, F. and Huang, J. Fault diagnosis based on measurement reconstruction of HPT exit pressure for turbofan engine, Chin. J. Aeronaut., 2019, 32, (5), pp 1156–1170.10.1016/j.cja.2019.03.032CrossRefGoogle Scholar
Bouzid, M.B. and Champenois, G. An efficient simplified physical faulty model of a permanent magnet synchronous generator dedicated to stator fault diagnosis part II: automatic stator fault diagnosis, IEEE. T. Ind. Appl., 2017, 53, (3), pp 2762–2771.10.1109/TIA.2017.2661841CrossRefGoogle Scholar
Cha, J., Ko, S., Park, S. and Jeong, E. Fault detection and diagnosis algorithms for transient state of an open-cycle liquid rocket engine using nonlinear Kalman filter methods, Acta. Astronaut., 2019, 163, pp 147–156.10.1016/j.actaastro.2019.03.075CrossRefGoogle Scholar
Li, Y.G. Gas turbine performance and health status estimation using adaptive gas path analysis. J. Eng. Gas. Turb. Power., 2010, 132, (4), p 041701.10.1115/1.3159378CrossRefGoogle Scholar
Fast, M., Palme´, T. and Genrup, M. A novel approach for gas turbine condition monitoring combining CUSUM technique and artificial neural network, Proceedings of the ASME Turbo Expo, Florida, USA, 2009.10.1115/GT2009-59402CrossRefGoogle Scholar
Zhou, D., Zhang, H. and Weng, S. A new gas path fault diagnostic method of gas turbine based on support vector machine, J. Eng. Gas. Turb. Power., 2015, 137, (10), p 102605.10.1115/1.4030277CrossRefGoogle Scholar
Zhao, L., Mo, C., Sun, T. and Huang, W. Aero engine gas-path fault diagnose based on multimodal deep neural networks, Wirel. Commun. Mob. Comput., 2020. 10.1155/2020/8891595CrossRefGoogle Scholar
Ying, S., Wang, B., Wang, L., Li, Q., Zhao, Y., Shang, J., Huang, H., Cheng, G., Yan, Z. and Geng, J. An improved KNN-based efficient log anomaly detection method with automatically labeled samples, ACM Trans. Knowl. Discov. Data., 2021, 15, (3), pp 1–22.10.1145/3441448CrossRefGoogle Scholar
Li, Y. and Guo, L. TCM-KNN scheme for network anomaly detection using feature-based optimizations, Acm Symposium on Applied Computing, Brazil, 2008. 10.1145/1363686.1364194CrossRefGoogle Scholar
Jain, P., Bajpai, M.S. and Pamula, R. A modified DBSCAN algorithm for anomaly detection in time-series data with seasonality, Int. Arab. J. Inf. Techn., 2022, 19, (1), pp 23–28.Google Scholar
Christopher, R.D.E., Song, S., Lee, H., Choi, D., Lim, J., Bok, K. and Yoo, J. Anomaly detection based on GCNs and DBSCAN in a large-scale graph, Electronics, 2024, 13, (13), p 2625.Google Scholar
Adesh, A., Shobha, G., Shetty, J. and Xu, L. Local outlier factor for anomaly detection in HPCC systems, J. Parallel. Distr. Com., 2024, 192, p 104923.10.1016/j.jpdc.2024.104923CrossRefGoogle Scholar
Pang, J., Pu, X. and Li, C. A hybrid algorithm incorporating vector quantization and one-class support vector machine for industrial anomaly detection, IEEE T. Ind. Inform., 2022, 18, (12), pp 8786–8796.10.1109/TII.2022.3145834CrossRefGoogle Scholar
Xu, D., Wang, Y., Meng, Y. and Zhang, Z. An improved data anomaly detection method based on isolation forest, 10th International Symposium on Computational Intelligence and Design, Hangzhou, China, 2017. 10.1109/ISCID.2017.202CrossRefGoogle Scholar
Li, N., Liu, X., Liu, Z., Mao, L., Zhao, L. and Wang, X. Anomaly detection in power grid IoT system based on isolated forest, IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology, Melbourne, Australia, 2022. 10.1145/3498851.3498922CrossRefGoogle Scholar
Sakurada, M. and Yairi, T. Anomaly detection using autoencoders with nonlinear dimensionality reduction, Proceedings of the MLSDA 2014 2nd Workshop on Machine Learning for Sensory Data Analysis, Australia, 2014.10.1145/2689746.2689747CrossRefGoogle Scholar
Zhou, C., Paffenroth, R.C. Anomaly detection with robust deep autoencoders, Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax, Canada, 2017. 10.1145/3097983.3098052CrossRefGoogle Scholar
Han, Y., Ma, Y., Wang, J. and Wang, J. Research on ensemble model of anomaly detection based on autoencoder, 2020 IEEE 20th International Conference on Software Quality, Reliability and Security, Macau, China, 2020.10.1109/QRS51102.2020.00060CrossRefGoogle Scholar
Song, J., Kim, K., Oh, J., Cho, S., Oh, A., Neumann, T., Globerson, A., Saenko, K., Hardt, M. and Levine, S. Memto: memory-guided transformer for multivariate time series anomaly detection, Advances in Neural Information Processing Systems, New Orleans, USA, 2023. Google Scholar
Fentaye, A.D. and Kyprianidis, K.G. Gas turbine prognostics via Temporal Fusion Transformer, Aeronaut. J., 2024, 128, (1325), pp 1594–1609.10.1017/aer.2024.40CrossRefGoogle Scholar
Miao, Z., Chen, G., Liao, X., Dai, J. and He, Y. Self-supervised dual-layer 2D normalizing flow method for industrial anomaly detection, Appl. Soft. Comput., 2024, 163, p 111928.10.1016/j.asoc.2024.111928CrossRefGoogle Scholar
Dai, E. and Chen, J. Graph-augmented normalizing flows for anomaly detection of multiple time series, arXiv preprint, 2022.Google Scholar
Zhang, X., Wang, C., Zhou, W., Xu, J. and Han, T. Trustworthy diagnostics with out-of-distribution detection: a novel max-consistency and min-similarity guided deep ensembles for uncertainty estimation, IEEE Internet. Things. J., 2024, 11, (13), pp 23055–23067.10.1109/JIOT.2024.3387481CrossRefGoogle Scholar
Han, X., Liu, Y., Zhang, Z., Lü, X. and Li, Y. Sparse auto-encoder combined with kernel for network attack detection, Comput. Commun., 2021, 173, pp 14–20.10.1016/j.comcom.2021.03.004CrossRefGoogle Scholar
Lin, Y., Liang, Y., Yoshida, S., Feng, X. and Guan, R. A hybrid algorithm of extreme learning machine and sparse auto-encoder, International Conference on Smart Computing and Communication, Shenzhen, China, 2016.10.1007/978-3-319-52015-5_20CrossRefGoogle Scholar
San Martin, G., Droguett, E.L., Meruane, V. and Moura, M. Deep variational auto-encoders: a promising tool for dimensionality reduction and ball bearing elements fault diagnosis, Struct. Health. Monit., 2019, 18, (4), pp 1092–1128.10.1177/1475921718788299CrossRefGoogle Scholar
Vincent, P., Larochelle, H., Lajoie, I., Bengio, Y. and Manzagol, P. Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion, J. Mach. Learn. Res., 2010, 11, pp 3371–3408.Google Scholar
Hu, D., Zhang, C., Yang, T. and Fang, Q. A deep autoencoder with structured latent space for process monitoring and anomaly detection in coal-fired power units, Reliab. Eng. Syst. Saf., 2025, 261, p 111060.10.1016/j.ress.2025.111060CrossRefGoogle Scholar
Wang, X. and Chen, Y. Unsupervised detection of multivariate geochemical anomalies using a high-performance deep autoencoder Gaussian mixture model, J. Geochem. Explor., 2025, 271, p 107671.10.1016/j.gexplo.2025.107671CrossRefGoogle Scholar
Hooshmand, M.K., Huchaiah, M.D., Alzighaibi, A.R., Hashim, H., Atlam, E. and Gad, I. Robust network anomaly detection using ensemble learning approach and explainable artificial intelligence (XAI), Alex. Eng. J., 2024, 94, pp 120–130.10.1016/j.aej.2024.03.041CrossRefGoogle Scholar
Lin, S., Zhang, M., Cheng, X., Shi, L., Gamba, P. and Wang, H. Dynamic low-rank and sparse priors constrained deep autoencoders for hyperspectral anomaly detection, IEEE Trans. Instrum. Meas., 2024, 73, pp 1–18.Google Scholar
Ding, K., Li, J., Bhanushali, R. and Liu, H. Deep anomaly detection on attributed networks, Proceedings of the 2019 SIAM International Conference on Data Mining, Calgary, Canada, 2019.10.1137/1.9781611975673.67CrossRefGoogle Scholar
Xu, D., Ricci, E., Yan, Y., Song, J. and Sebe, N. Learning deep representations of appearance and motion for anomalous event detection, British Machine Vision Conference, Swansea, UK, 2015.10.5244/C.29.8CrossRefGoogle Scholar
Moslemi, A. and Jamshidi, M. Unsupervised feature selection using sparse manifold learning: auto-encoder approach, Inf. Process. Manag., 2025, 62, (1), p 103923.10.1016/j.ipm.2024.103923CrossRefGoogle Scholar
Rifai, S., Vincent, P., Muller, X., Glorot, X. and Bengio, Y. Contractive auto-encoders: explicit invariance during feature extraction, Proceedings of the 28th International Conference on International Conference on Machine Learning, Bellevue, USA, 2011.10.1007/978-3-642-23783-6_41CrossRefGoogle Scholar
Pang, G., Shen, C., Cao, L. and Hengel, A. Deep learning for anomaly detection: a review, ACM Comput. Surv., 2021, 54, (2), pp 1–38.10.1145/3439950CrossRefGoogle Scholar
Lin, J. and Zhao, M. A review and strategy for the diagnosis of speed-varying machinery, 2014 International Conference on Prognostics and Health Management, Cheney, USA, 2014.10.1109/ICPHM.2014.7036368CrossRefGoogle Scholar
Fala, N., Georgalis, G. and Arzamani, N. Study on machine learning methods for general aviation flight phase identification, J. Aerosp. Inf. Syst., 2023, 20, (10), pp 636–647.Google Scholar
Yin, X., He, Q., Peng, S., Chen, D., Zhang, H. and Zhang, B. A health state prediction model for aeroengine based on multi-attribute belief rule base with considering monitoring error, Int. J. Fuzzy. Syst., 2025, 27, pp 1072–1095.10.1007/s40815-024-01808-xCrossRefGoogle Scholar
Fu, S., Zhong, S., Lin, L. and Zhao, M. A re-optimized deep auto-encoder for gas turbine unsupervised anomaly detection, Eng. Appl. Artif. Intell., 2021, 101, (12), p 104199.10.1016/j.engappai.2021.104199CrossRefGoogle Scholar
Vincent, P., Larochelle, H., Bengio, Y. and Manzagol, P. Extracting and composing robust features with denoising autoencoders, Proceedings of the 25th international conference on Machine learning, Helsinki, Finland, 2008. 10.1145/1390156.1390294CrossRefGoogle Scholar
Schneider, S., Antensteiner, D., Soukup, D. and Scheutz, M. Autoencoders - a comparative analysis in the realm of anomaly detection, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, New Orleans, USA, 2022.10.1109/CVPRW56347.2022.00216CrossRefGoogle Scholar
Ahasan, M.R., Haque, M.S., Akram, M.R., Momen, M.F. and Alam, M.G.R. Deep learning autoencoder based anomaly detection model on 4G network performance data, 2022 IEEE World AI IoT Congress, Seattle, USA, 2022.10.1109/AIIoT54504.2022.9817338CrossRefGoogle Scholar
Tong, C.D., Palazoglu, A. and Yan, X.F. An adaptive multimode process monitoring strategy based on mode clustering and mode unfolding, J. Process. Contr., 2013, 23, pp 1497–1507.10.1016/j.jprocont.2013.09.017CrossRefGoogle Scholar
Song, B., Ma, Y. and Shi, H. Multimode process monitoring using improved dynamic neighborhood preserving embedding, Chemom. Intell. Lab. Syst., 2014, 135, pp 17–30.10.1016/j.chemolab.2014.03.013CrossRefGoogle Scholar
Song, B., Yan, H., Shi, H. and Tan, S. Multisubspace elastic network for multimode quality-related process monitoring, IEEE Trans. Ind. Inf., 2019, 16, pp 5874–5883.10.1109/TII.2019.2959784CrossRefGoogle Scholar
Han, S., Hu, X., Huang, H., Jiang, M. and Zhao, Y. ADBench: anomaly detection benchmark, ArXiv, 2022.10.2139/ssrn.4266498CrossRefGoogle Scholar
Liu, B., He, J. and Liu, G. A study of working conditions based on cluster analysis, 2010 Asia-Pacific Power and Energy Engineering Conference, Chengdu, China, 2010.10.1109/APPEEC.2010.5448568CrossRefGoogle Scholar
Du, G., Meng, Q. and Yu, H. Research on modeling of cement denitrification under typical working conditions, 2018 IEEE 4th Information Technology and Mechatronics Engineering Conference, Chongqing, China, 2018.10.1109/ITOEC.2018.8740444CrossRefGoogle Scholar
Esteves, R.M., Hacker, T. and Rong, C. Competitive K-means, a new accurate and distributed K-means algorithm for large datasets, 2013 IEEE 5th International Conference on Cloud Computing Technology and Science, Bristol, UK, 2013.10.1109/CloudCom.2013.89CrossRefGoogle Scholar
Hinton, G.E. and Salakhutdinov, R.R. Reducing the dimensionality of data with neural networks, Science, 2006, 313, pp 504–507.10.1126/science.1127647CrossRefGoogle ScholarPubMed
Li, Z., Zhao, Y., Hu, X., Botta, N., Ionescu, C. and Chen, G.H. ECOD: unsupervised outlier detection using empirical cumulative distribution functions, IEEE Trans. Knowl. Data. Eng., 2022, 35, (12), pp 12181–12193.10.1109/TKDE.2022.3159580CrossRefGoogle Scholar
Continental Aerospace Technologies GmbH. Operation & Maintenance Manual, 2020. Google Scholar
Tuli, S., Casale, G. and Jennings, N. TranAD: deep transformer networks for anomaly detection in multivariate time series data, Proc. VLDB Endow., 2022, 15, (6), pp 1201–1214.10.14778/3514061.3514067CrossRefGoogle Scholar
Audibert, J., Michiardi, P., Guyard, F., Marti, S. and Zuluaga, M.A. USAD: unsupervised anomaly detection on multivariate time series, Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, USA, 2020. Google Scholar
Li, N., Yang, J., Zhou, R. and Liang, C. Determination of knock characteristics in spark ignition engines: an approach based on ensemble empirical mode decomposition, Meas. Sci. Technol., 2016, 27, (4), p 045109.10.1088/0957-0233/27/4/045109CrossRefGoogle Scholar
Longwic, R., Sen, A.K., Górski, K., Lotko, W. and Litak, G. Cycle-to-cycle variation of the combustion process in a diesel engine powered by different fuels, J. Vibroeng., 2011, 13, (1), pp 120–127.Google Scholar
Verma, R. and Ganguli, R. Denoising jet engine gas path measurements using nonlinear filters, IEEE/ASME Trans. Mechatron., 2005, 10, (4), pp 461–464.10.1109/TMECH.2005.852454CrossRefGoogle Scholar