Implementation of an Edge-IoT-Based Real-Time Predictive Maintenance and VOCs Monitoring and Control Service Platform for the Safe Operation of Hybrid RTO Systems : Solution Service for Identifying Failure Causes and Establishing Safety Measures
This study aims to implement an Edge-IoT-based real-time predictive maintenance service platform to effectively support the full life cycle of maintenance, repair, and operation(MRO) for a hybrid regenerative thermal oxidizer(RTO). In distributed operating environments where remote RTO facilities and centralized monitoring systems are spatially separated, continuous condition monitoring, early identification of abnormal signs, and systematic response procedures are essential. Accordingly, this study seeks to enhance operational safety and maintenance efficiency by structuring failure types, root causes, and response procedures based on actual failure cases and representative operating scenarios of RTO facilities. To achieve this objective, an integrated platform was designed and implemented, consisting of field sensors, Edge-IoT gateway devices, an MQTT-based data collector, Redis and MySQL databases, a web server, a hybrid application, and a monitoring module. In addition, an analysis procedure considering the application of low-power edge devices was designed to enable independent condition assessment even under resource-constrained field environments. The target data included time-series variables reflecting both RTO operating conditions and emission characteristics, such as temperature, pressure, flow rate, current, vibration, and volatile organic compound(VOC) concentration. These data were preprocessed and utilized for abnormality detection and failure prediction. The implementation results confirmed that the proposed platform can collect and visualize the operating status of remote RTO facilities in near real time. Furthermore, when abnormal operating conditions occur, the platform supports rapid maintenance decision-making by providing alarms, storing historical records, and linking predefined response procedures. In particular, by integrating VOC emission characteristics with RTO operating conditions, the platform demonstrated its potential not only for improving facility safety but also for supporting emission reduction and environmental management. Moreover, the integration of Edge-IoT, lightweight AI-based analysis, and 3D digital twin-based monitoring represents a distinctive contribution compared with conventional remote monitoring systems that primarily focus on simple supervisory functions. In conclusion, this study presents a practical reference model for the digital transformation of RTO facilities and the establishment of a predictive maintenance framework. The proposed platform enables active condition assessment under limited field infrastructure and supports systematic responses to abnormal signs, thereby contributing to improved maintenance efficiency, operational safety, and environmental management capability. If further applied and expanded to various industrial RTO facilities, the proposed approach is expected to promote data-driven safety management and the wider adoption of predictive maintenance-oriented operational practices.