The 12th International Conference on Indoor Air Quality, Ventilation & Energy Conservation in Buildings(IAQVEC 2026), May 18-22, University of Southern California Los Angeles, California, USA
Environment|구충완
Toward Carbon-Neutral Construction: Clustering-Based Benchmarking Framework for Site Electricity Usage and Anomalous Usage Detection
2026International Conference
Research AreaEnvironment
Professor구충완
AuthorsSeo, S., Choi, D., Choi, Y., Ahn, S., Lee, W., and Koo, C. (Corresponding Author)
Publication Year2026
Journal (Volume, Issue, Pages)The 12th International Conference on Indoor Air Quality, Ventilation & Energy Conservation in Buildings(IAQVEC 2026), May 18-22, University of Southern California Los Angeles, California, USA
AbstractIn response to the global climate crisis, the construction industry must establish actionable greenhouse gas (GHG) reduction strategies to achieve carbon neutrality by 2050. However, current site-level energy management—relying mainly on monthly billing—fails to capture temporal and spatial patterns of electricity consumption (Scope 2), limiting the identification of inefficiencies. This study proposes a data-driven framework leveraging high-resolution electricity data to diagnose energy use at construction sites with precision.
Approximately 16 months of 15-minute interval data were collected from Korean residential building projects and analyzed by distinguishing between construction site facilities and site offices. Time-series clustering was applied to extract representative consumption patterns, while calendar-based energy maps visualized seasonal and operational trends. Four distinct usage patterns were identified, with anomalies such as unnecessary weekend cooling of site facilities in summer and standby heating in site offices during winter. The proposed framework enhances visibility into real-time consumption patterns, enabling data-driven decision-making and targeted energy-saving interventions. These findings support the transition from reactive to proactive energy management in construction and provide a foundational tool for achieving national GHG reduction targets.
Approximately 16 months of 15-minute interval data were collected from Korean residential building projects and analyzed by distinguishing between construction site facilities and site offices. Time-series clustering was applied to extract representative consumption patterns, while calendar-based energy maps visualized seasonal and operational trends. Four distinct usage patterns were identified, with anomalies such as unnecessary weekend cooling of site facilities in summer and standby heating in site offices during winter. The proposed framework enhances visibility into real-time consumption patterns, enabling data-driven decision-making and targeted energy-saving interventions. These findings support the transition from reactive to proactive energy management in construction and provide a foundational tool for achieving national GHG reduction targets.
Site Electricity ConsumptionTime-Series ClusteringAnomaly DetectionEnergy Calendar MapsScope 2 Emissions