Implementing an Edge-IoT System for Real-Time Information Gathering of RTOs : Implementing Architecture for IoT Data Collection
,Chi Young Jang,Dal Hwan Yoon,Myung Kee Jang,Jae Wan Jeong,Ji Heon Kim,Min Woo Kim,Hoon Min Park,Tae Young Lim
DOI: https://doi.org/10.52902/kjsc.2024.34.33
2024-09-30
Forum of Public Safety and Culture
Abstract:In this study, an IoT architecture is implemented to collect and determine real-time status information for remote maintenance of heat storage thermal oxidizer. IoT architecture systems interface with PLC (Programmable Logic Control), which controls RTO operation monitoring, and transmit real-time data to remote servers, and channel data received through IoT is analyzed to build a database and visualize the characteristics of RTO devices. IoT architecture implementation for real-time monitoring, failure determination, and maintenance of heat storage combustion oxidation facilities interface IoT circuitry to the PLC control panel that controls monitoring. Compatibility and information security are considered to facilitate maintenance at a distance. The received data is segmented for each channel and analyzed with a visualization algorithm to detect abnormality in segmentation data. At this time, the visualized data are judged as normal, insufficient, or warning according to the threshold value. The common software stack used in existing IoT systems is necessary for computing, portability, and ease of management that allows data processing to be moved between the edge and the cloud. When an edge node interacts with a specific cloud backend in the monitoring PLC of multiple RTO devices, the network bandwidth between the edge and the cloud presents a bottleneck of large-scale data transfer. In addition, in a clustered system of many nodes where data is transferred between the edge and the cloud, data can be lost forever due to a malfunction of a single edge node. Especially in the case of node failure or intermittent long-distance network connection problems, it is necessary to implement local fault tolerance to preserve system state locally at the edge. To address this problem, this paper proposes CEFIoT, a new fault-tolerant architecture for IoT applications by adopting state-of-the-art cloud technology and also deploying it to edge computing. The CEFIoT architecture consists of three layers: (i) Application Isolation, (ii) Data Transport, and (iii) Multi-Cluster Management layer. Based on this tiered design, the architecture allows computing deployments on edges or clouds without source code modifications. In addition to real-time status data for each RTO facility, information management on the histories of individual devices and parts is required, and facility data according to abnormal situations are essential for preservation through time series data learning. However, since it can be difficult to reproduce each abnormal situation, it is necessary to virtually reproduce the abnormal situation through facility modeling through normal state data.