It is of paramount importance to provide Indoor Air Quality (IAQ) as it can improve health, productivity, and well-being of the users while reducing the energy consumption. Needless to mention that the building sector consumes nearly 40% global energy. IAQ is fundamentally dependent on the strategic placement of diffusers and exhausts, and a critical configuration for the placement of the valves can profoundly impact ventilation effectiveness, occupant comfort, and energy consumption. Traditional designs rely on simulations with predefined locations for the inlets and outlets, and static sensor locations, often failing to capture the complex, three-dimensional nature of indoor air distribution. This can lead to a performance gap between design intent and real-world operation. To rectify this predicament, this paper introduces a novel methodology that integrates a dynamic digital twin, autonomous robotic sensing, and uncertainty quantification (UQ) to identify robustly optimal ventilation configurations. The method utilizes a simulation platform to create a digital twin of the experimental setup equipped with air inlets and outlets. A mobile robotic agent then systematically establishes a baseline for the digital twin by autonomously navigating the space to generate high-resolution, spatio-temporal maps of thermal and air quality parameters for various inlet and outlet location scenarios. These rich datasets are used not for deterministic optimization, but to drive a UQ analysis that characterizes the performance robustness of each configuration against key real-world uncertainties, such as sensor error and fluctuations in supply airflow. The best layout was found to be sensitive to operational variance, whereas a UQ-informed robust layout yielded comparable average IAQ. This framework provides a new paradigm for the evidence-based design and virtual commissioning of ventilation systems. It ensures that selected configurations are not only efficient but also robust and reliable under the dynamic conditions of actual building operation.
Xiao et al. (Tue,) studied this question.