基于双注意力LSTM-AdvAE的无监督异常检测与预警——面向多维时间序列的工业故障智能诊断方法
Unsupervised Anomaly Detection and Early Warning Based on Dual-Attention LSTM-AdvAE—Intelligent Industrial Fault Diagnosis Method for Multidimensional Time Series
摘要: 无监督异常检测在信息科学与工业监测中具有重要应用价值，但现有方法在多维时间序列的时空特征建模、微弱故障识别等方面仍存在不足。本文提出一种基于双注意力LSTM‑AdvAE的无监督异常检测与预警模型，通过TFDAM双注意力模块对特征与时间维度进行加权增强，结合LSTM捕捉时序依赖关系，并利用改进对抗自编码器实现更鲁棒的异常识别。在公开数据集上的实验表明，该方法在精确率、召回率及F1分数上均优于现有主流方法，能够有效实现工业设备故障的早期预警。
Abstract: Unsupervised anomaly detection (UAD) has been widely used in industrial monitoring and information science. However, current methods hardly model spatio-temporal dependencies and detect weak faults effectively for high-dimensional time series. This paper proposes a Dual-Attention LSTM-AdvAE model for unsupervised anomaly detection and early warning. The model uses a TFDAM module to weight features and temporal dimensions, adopts LSTM to capture long-term dependencies, and applies an improved adversarial autoencoder for robust anomaly scoring. Experiments on public datasets demonstrate that this method outperforms existing mainstream approaches in precision, recall and F1-score, enabling effective early warning of industrial equipment faults.
文章引用：张之翼, 杨卫华. 基于双注意力LSTM-AdvAE的无监督异常检测与预警——面向多维时间序列的工业故障智能诊断方法[J]. 应用数学进展, 2026, 15(4): 576-584. https://doi.org/10.12677/aam.2026.154184

参考文献

[1] Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986) Learning Representations by Back-Propagating Errors. Nature, 323, 533-536. [Google Scholar] [CrossRef] 
[2] Goodfellow, I., Pouget-Abadie, J., Mirza, M., et al. (2014) Generative Adversarial Nets. NIPS’14: Proceedings of the 28th International Conference on Neural Information Processing Systems, Volume 2, 2672-2680.
[3] Shi, Y., Lei, M., Ma, R. and Niu, L. (2019) Learning Robust Auto-Encoders with Regularizer for Linearity and Sparsity. IEEE Access, 7, 17195-17206. [Google Scholar] [CrossRef] 
[4] Zong, B., Song, Q., Min, M.R., et al. (2018) Deep Autoencoding Gaussian Mixture Model for Unsupervised Anomaly Detection. International Conference on Learning Representations, Vancouver, 30 April-3 May 2018.
[5] Lin, S., Clark, R., Birke, R., Schonborn, S., Trigoni, N. and Roberts, S. (2020) Anomaly Detection for Time Series Using VAE-LSTM Hybrid Model. 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona, 4-8 May 2020, 4322-4326. [Google Scholar] [CrossRef] 
[6] Su, Y., Zhao, Y., Niu, C., Liu, R., Sun, W. and Pei, D. (2019) Robust Anomaly Detection for Multivariate Time Series through Stochastic Recurrent Neural Network. Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, 4-8 August 2019, 2828-2837. [Google Scholar] [CrossRef] 
[7] Audibert, J., Michiardi, P., Guyard, F., Marti, S. and Zuluaga, M.A. (2020) USAD: Unsupervised Anomaly Detection on Multivariate Time Series. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 6-10 July 2020, 3395-3404. [Google Scholar] [CrossRef] 
[8] Khan, W., Mohd, A., Suaib, M., Ishrat, M., Shaikh, A.A. and Faisal, S.M. (2025) Residual-Enhanced Graph Convolutional Networks with Hypersphere Mapping for Anomaly Detection in Attributed Networks. Data Science and Management, 8, 137-146. [Google Scholar] [CrossRef] 
[9] Khan, W., Abidin, S., Arif, M., Ishrat, M., Haleem, M., Shaikh, A.A., et al. (2024) Anomalous Node Detection in Attributed Social Networks Using Dual Variational Autoencoder with Generative Adversarial Networks. Data Science and Management, 7, 89-98. [Google Scholar] [CrossRef] 
[10] Ergen, T. and Kozat, S.S. (2020) Unsupervised Anomaly Detection with LSTM Neural Networks. IEEE Transactions on Neural Networks and Learning Systems, 31, 3127-3141. [Google Scholar] [CrossRef] [PubMed]

为你推荐