Formal calibration methodology relating to CFD models of naturally ventilated internal environments

Abstract

The building sector is responsible for 40% of total global energy consumption. Half of the energy consumed by buildings relates to heating, ventilation and air conditioning (HVAC) systems. Natural ventilation can provide a sustainable solution aimed at maintaining healthy and comfortable environmental conditions in buildings. However, the effective design and operation of new and retrofitted buildings that incorporate natural ventilation systems is often complex and requires the utilisation of accredited simulation tools and reliable field measurements. The Network Embedded Systems (NEMBES) project is a Higher Education Authority, Ireland (HEA) funded inter-institutional and multi-disciplinary research program that examines the application of network embedded systems in the design and management of built environments, including the health, facilities management and transportation sectors. This research forms a core part of the NEMBES project through the development of reliable datasets that could facilitate (i) the specific design, deployment and management of novel wireless sensor networks (WSNs) and (ii) distributed control technologies in naturally ventilated buildings, as part of recognised facilities management activities in sustainable buildings. This research specifically focuses on the development of a formal systematic calibration methodology for computational fluid dynamics (CFD) models utilising field measurements in operating buildings. Published literature to date lacks explicit and documented methodologies for calibrating CFD models, in order to develop robust computational models that accurately represent environmental conditions in operating naturally ventilated buildings. Despite existing best practice guidelines for verification and validation of CFD models, there is a need for comprehensive studies that systematically guide, explain and assess simulated data and post-occupancy field measurements in naturally ventilated built environments. Thus, the motivation for calibrated CFD models of indoor environments is to examine the possibility of employing CFD to simulate, and potentially operate, naturally ventilated environments, considering the uncertainty of available data. In this research, calibrated 3D virtual models of two internal environments were developed. The CFD models were supported by physical field measurements from wireless sensors and a locally installed weather station. The numerical boundary conditions were analysed for their effect on model results and its accuracy, by utilising uncertainty and regression analysis. This provided a systematic and robust calibration process of CFD models that represented environmental conditions in operating buildings.