Performance Evaluation of Smart Home Digital Twins: A Thermal Comfort Case Study

Abstract

Decarbonising residential buildings requires control strategies that can maintain thermal comfort while minimising energy use. Traditional rule‑based HVAC controls often fall short, as they struggle to respond to rapidly changing indoor conditions and evolving energy constraints. This study presents a real‑time Digital Twin (DT) framework for optimising thermal comfort in smart homes, combining low‑cost IoT sensing, CFD simulation (ANSYS Fluent), and dynamic control algorithms implemented in MATLAB. The framework was validated in an occupied residential setting using a structured Key Performance Indicator (KPI) approach across eight scenarios with varying window apertures (0-0.3 m) under passive and active heating and ventilation modes. Performance was assessed using ISO 7730 comfort metrics (PMV, PPD), statistical accuracy indicators (MAE, RMSE, R²), and operational feasibility measures (latency, spatial consistency, cost efficiency). The system achieved PMV targets in 88% of cases, with full compliance (<10% PPD). Passive ventilation scenarios showed the strongest comfort performance and favourable practical deployment characteristics, while energy-related outcomes remained scenario-dependent. Predictive reliability was supported by Wilcoxon signed‑rank tests (p > 0.05 in 53% of cases). Overall, the proposed KPI-driven DT framework offers an occupant-responsive and scalable solution for real-time residential thermal comfort management, while also clarifying the conditions required for broader energy benefits to be realised.

Conflict of Interest Statement

The authors declare no conflicts of interest.