The rapid development of smart factories, combined with the increasing complexity of production equipment, has resulted in a large number of multivariate time series that can be recorded using sensors during the manufacturing process. The anomalous patterns of industrial production may be hidden by these time series. Previous LSTM-based and machine-learning-based approaches have made fruitful progress in anomaly detection. However, these multivariate time series anomaly detection algorithms do not take into account the correlation and time dependence between the sequences. In this study, we proposed a new algorithm framework, namely, graph attention network and temporal convolutional network for multivariate time series anomaly detection (GTAD), to address this problem. Specifically, we first utilized temporal convolutional networks, including causal convolution and dilated convolution, to capture temporal dependencies, and then used graph neural networks to obtain correlations between sensors. Finally, we conducted sufficient experiments on three public benchmark datasets, and the results showed that the proposed method outperformed the baseline method, achieving detection results with F1 scores higher than 95% on all datasets.

Recent unsupervised framework-based anomaly detection methods in Internet of Things (IoT) have been proved successful by learning temporal pattern representations from normal time series data. However, such methods have long suffered from weakened normal-abnormal boundary incurred by discrimination-insufficient temporal pattern representations, due to: (1) normal time series often suffer from some unprocessed abnormal noise, interfering with model’s discriminative representations between normal and abnormal points. (2) Previous methods rarely focus on contextual semantic features of time series that are also crucial for constructing discriminative representations. In this paper, we propose MPFormer, a transformer-based multi-patch contrastive learning framework. This novel framework leverages contrastive learning to guide discriminative feature learning. It incorporates a data augmentation strategy to generate positive samples and maximizes the similarity between positive sample pairs, thereby capturing inherent representations of normal temporal patterns and enhancing robust discrimination ability to abnormal noise. To further facilitate discriminative representation modeling, we divide input time series into multiple patches, and design a transformer-based dual attention module to explore contextual semantic features in both inter-patch and intra-patch views. Benefiting from discriminative representations being modeled, MPFormer effectively strengthens the normal-abnormal boundary, thus improving detection accuracy. Experimental results on six widely adopted datasets demonstrate that our proposed MPFormer outperforms existing baseline methods.

The mixing process of battery production contains a variety of monitoring sensors. These sensors generate a large number of multivariate time series during production, which reflect the potential operating conditions of the production equipment. By accurately detecting anomalies in the equipment, battery quality can be improved. However, due to the diverse types of anomalies in the mixing process and the large number of subtle anomalies, it is challenging to construct a robust and accurate anomaly detection model. In this paper, we propose TPAD, a multivariate time series anomaly detection model based on temporal pattern, which learns serial temporal pattern based on attention mechanism and detects anomalies by temporal pattern discrepancy, making the model adaptive to detect multiple types of anomalies. Moreover, the model improves the accuracy of detecting subtle anomalies through an adversarial training reconstruction method. Finally, we conducted experiments on two public datasets and a battery mixing process dataset from a battery factory, and the experimental results outperformed the baseline methods, with an average F1 score of over 92% on the three datasets.