ScholarWorks@성균관대학교

초록

The digital twin is a virtual replica of a physical system, overcoming spatial limitations of processes challenging for direct observation, and is widely used for real-time process monitoring. Recently, the integration of digital twins with internet of things (IoT) and information and communication technologies (ICT) has increased the demand for real-time data processing capabilities. The addition of various sensors and devices has led to the challenge of effectively processing big data, especially unstructured data such as images and logs. Existing industrial communication protocols struggle to transmit and process large-scale unstructured data in digital twins integrated with IoT and ICT technologies, resulting in limitations such as network overload, data reliability, and lack of system scalability. To overcome these limitations, this study developed a digital twin system that can process unstructured data based on a distributed network and applied it to implement a system that can monitor the directed energy deposition (DED) process in real time. A testbed for process monitoring was created, and an environment was made up to collect unstructured data from sensors and devices in real time. A distributed network was built using Kafka, a publish-subscribe model, as a message broker to process unstructured data. Data was collected and pre-processed in edge computing systems, and computation-intensive tasks were performed in cloud computing systems. Subsequently, the network system was optimized to enhance real-time performance, communication efficiency, and scalability. It was then applied to the DED digital twin, confirming its effectiveness in real-time process monitoring with unstructured data. Performance evaluation showed that the optimized system minimized latency and reduced bandwidth consumption, ensuring stable real-time data processing under high-load conditions. The system achieved an average latency of 26.62 ms for IR camera data and 48.89 ms for CCD camera data, significantly improving responsiveness and network stability.

키워드