Multi-stage calibration framework for a digital twin model in building operations: Cold chain logistics centers case study

Abstract

This paper presents a multi-stage calibration framework for digital twins in building operations for cold chain logistics centers, focusing on key aspects such as temperature dynamics, cooling loads, and power consumption during such building operations. The rapid expansion of cold chain logistics centers has introduced significant challenges in ensuring product quality, optimizing energy consumption, and reducing operational costs. Digital twin-enabled building operations offer a potential solution to address these challenges. The proposed building digital twin, developed using EnergyPlus and Python, integrates sensor data with particle swarm optimization (PSO) algorithms to systematically calibrate key parameters such as internal thermal mass, air infiltration, and HVAC performance. Calibration is performed with a time step of one-minute, improving model accuracy by capturing transient dynamics that often overlooked by conventional hourly calibration methods. A real-world building was used to validate the proposed building digital twin model structure and calibration framework. Experimental results demonstrated the ability of the digital twin to predict building operating temperatures and energy consumption with high accuracy. The study highlights the benefits of using temperature and power sensor data as the primary inputs for model calibration, showing the potential on reducing reliance on more complex and intrusive measurement techniques. Furthermore, a multi-objective particle swarm optimization (MOPSO) algorithm was implemented to further verify the theoretical feasibility of the proposed multi-stage calibration framework.