A user-centric temperature sensor deployment method under digital twin leveraging occupancy information

The temperature sensor deployment for indoor temperature field estimation is critical in managing thermal comfort and achieving energy savings in the digital twin (DT) of building energy systems. Current temperature sensor deployment methods prioritize constructing the true temperature field, which yield increasing cost and added DT computing dimensions. Additionally, they frequently overlooked the significance of information from users' actual usage scenarios for sensor deployment and failed to balance multiple aspects like cost, coverage, and user satisfaction from a user-centric perspective. To solve these issues, a practical, low-cost, and user-centric temperature sensor deployment method leveraging occupancy information is proposed. Based on the first principle, it introduces user information directly from the perspective of improving user satisfaction for deploying temperature sensors. The user-centric multi-objectives temperature sensor deployment model is established by coverage model based on wireless fidelity (WiFi) connection points and user satisfaction metrics. To solve the multi-objective optimization problem with multivariable, constraints, and nonlinear, the Improved Multi-Objective Particle Swarm Optimization (IMOPSO) algorithm is developed. Results show that the proposed method increases WiFi-based coverage and satisfaction metrics. The room-level indoor temperature could be accurately estimated with a steady-state error of 0.199 (average RMSE) and dynamic-state error of 0.298 (heating, average RMSE) based on the proposed deployment method. These results demonstrate that the proposed user-centric approach provides a novel and practical solution for air temperature sensor deployment.