Dr. Sungmin Yoon’s research team from the School of Civil Architectural Engineering and Landscape Architecture has developed a virtual sensing technology for building operational phases that autonomously evolves and self-calibrates within a digital twin environment.

Conventional virtual sensor technology has primarily focused on products in the manufacturing industry, utilizing controlled laboratory settings for research and development. However, in the building sector, where “every building is different,” the architectural uniqueness and complex systems make it challenging to establish laboratory environments and develop high-performance virtual sensors. Particularly, there is a technical contradiction when physical sensors are required to develop virtual sensors for specific variables.

This study overcomes these limitations by creating an innovative modeling technique and algorithm that enables virtual sensors to be autonomously generated, expanded, and continuously self-calibrated within actual building digital twin environments beyond the lab. This technology implements the concept of in-situ virtual sensing through nonintrusive indirect modeling, which integrates physical theories of building systems with surrounding sensor data without relying on direct measurements of physical variables.

<Digital Twin Living Lab and Winter District Heating System Validation Results>

Validation results demonstrated the performance of this technology, achieving an RMSE of 0.27°C for hot water temperature in district heating systems in residential complexes and an annual MAPE of less than 1.5% for flow rates in centralized HVAC systems over a three-month winter period.

<In-situ Virtual Sensing in Building Digital Twins: Autonomous Sensor Creation, Expansion, and Continuous Calibration in Building Operations>

Dr. Yoon stated, "This field-based virtual sensing technology can potentially replace all thermometers and flow meters commonly used in the return lines of HVAC systems, significantly reducing sensor installation and maintenance costs in large-scale HVAC systems for plants, semiconductor clusters, communities, and campuses." He also mentioned ongoing development of virtual sensing-based AI operational management technology using GPT agents within the SKKU-SSIT program to explore its applicability in semiconductor clusters.

<Urban Building Operational Management Platform with Virtual Sensing: T-ranno>

This research was supported by the National Research Foundation of Korea’s Basic Research Program and published in Journal of Industrial Information Integration (Top JCR category, December 2023).

- Title: In situ virtual sensors in building digital twins: framework and methodology

- Authors:Sungmin Yoon, Jaebum Koo, Youngwoong Choi (Sungkyunkwan University)

- Journal: Journal of Industrial Information Integration (IF 15.7, as of publication year)

- DOI: https://doi.org/10.1016/j.jii.2023.100532