In industry, sensor-based monitoring of equipment or environment has become a necessity. Instead of using a single sensor, multi-sensor system is used to fully detect abnormalities in complex scenarios. Recently, physical models, signal processing technology, and various machine learning models have improved the performance. However, these methods either do not consider the potential correlation between features or do not take advantage of the sequential changes of correlation while constructing an anomaly detection model. This paper firstly analyzes the correlation characteristic of a multi-sensor system, which shows a lot of clues to the anomaly/fault propagation. Then, a multi-sensor anomaly detection method, which finds and uses the correlation between features contained in the multidimensional time-series data, is proposed. The method converts the multidimensional time-series data into temporal correlation graphs according to time window. By transforming time-series data into graph structure, the task of anomaly detection is considered as a graph classification problem. Moreover, based on the stability and dynamics of the correlation between features, a structure-sensitive graph neural network is used to establish the anomaly detection model, which is used to discover anomalies from multi-sensor system. Experiments on three real-world industrial multi-sensor systems with anomalies indicate that the method obtained better performance than baseline methods, with the mean value of F1 score reaching more than 0.90 and the mean value of AUC score reaching more than 0.95. That is, the method can effectively detect anomalies of multidimensional time series.

The power transformer is an example of the key equipment of power grid, and its potential faults limit the system availability and the enterprise security. However, fault prediction for power transformers has its limitations in low data quality, binary classification effect, and small sample learning. We propose a method for fault prediction for power transformers based on dissolved gas chromatography data: after data preprocessing of defective raw data, fault classification is performed based on the predictive regression results. Here, Mish-SN Temporal Convolutional Network (MSTCN) is introduced to improve the accuracy during the regression step. Several experiments are conducted using data set from China State Grid. The discussion of the results of experiments is provided.

Purpose: Remaining useful life (RUL) prediction for power transformer maintenance is a challenging task on heterogeneous data. Monitoring data of power transformers are not always compatible or in an identical format; therefore, RUL predictions traditionally work separately on different data. Moreover, chemical molecules used in RUL prediction can be transformed into each other under different conditions, thus forming a complete graph with uncertain adjacency matrix (UAM). This study aims to find and evaluate a new model to achieve better results of RUL prediction than the other baselines. Design/methodology/approach: In this work, the authors propose a spatiotemporal complete graph convolutional network (STCGCN) for RUL prediction in two branches, in which daily and hourly features are extracted from correlated heterogeneous data separately. This study provides a thorough evaluation of the proposed model on real-world data and compare the proposed model with state-of-the-art RUL prediction models. Findings: By using the multibranch structure and EucCos similarity aggregation, STCGCN was able to capture the dynamic spatiotemporal patterns on a variety of heterogeneous data and obtain more accurate prediction results, compared to other time series prediction methods. Originality/value: In this work, the authors propose a novel multibranch structure to compute feature maps from two heterogeneous data sources efficiently and a novel similarity aggregation method to compute the spatial UAM within the complete graph. Compared with traditional time series prediction models, the model pays attention to the spatial relationships in time series data.

In light of the rapid advancement of technologies related to the Internet of things (IoT), IoT service platforms have become one of the main solutions for providing intelligent and efficient services in the industrial sector. Scheduling is an effective means to match resources and guarantee quality of service. However, existing service scheduling models and methods have not fully considered the special needs of new IoT platforms. Therefore, this article summarizes the special requirements of the new IoT platform, including the heterogeneity of IoT service platform resources, complexity and diversity of tasks, as well as considering the demand for low energy consumption and low latency. Constructed a multi-granularity task scheduling model for cloud-edge collaborative environments, which takes the special needs mentioned above into account. Combined with priority experience replay and importance sampling, a task scheduling algorithm priority replay with importance-based method in Actor Critic (PRIME-AC) based on deep reinforcement learning is proposed. The experimental results show that PRIME-AC has better performance in both task execution delay and load balancing than other baselines.

With the development of technology, mobile robots are increasingly deployed in real‐world environments. To enable robots to work safely in a variety of terrain environments, we proposed a ground‐type detection method based on the Hilbert–Huang transform (HHT) and attention‐based spatiotemporal coupled network. Taking a dataset containing multiple sets of robot signals from a Kaggle competition as an example; we use the proposed method to classify the signals and thus achieve a terrain classification of the robot’s location. Firstly, the signal data were processed using the discrete wavelet transform for noise reduction, and all channels in the dataset were ranked by importance using the permutation importance method. Next, the instantaneous frequencies of the two most important channels were extracted using the HHT and added to the original dataset to expand the feature dimension. Then the features in the expanded dataset were extracted by the convolutional neural network, long short‐term memory, and attention module. Afterward, the fully extracted features were passed into the fully connected layer for classification, and an average classification accuracy of 83.14% was obtained. The effectiveness of each part in our method was demonstrated using ablation experiments. Finally, we compared our method with some common methods in the field and found that our method obtained the highest classification accuracy, proving the superiority of the proposed method.