| Popis: |
Buildings are constructed to provide humans with a safe and comfortable environment to sleep, work and socialise. Therefore, they must be designed and operated to satisfy occupants’ needs and requirements without being resource-intensive and harmful to the environment. However, a significant obstacle in realising increased occupant satisfaction and high-performing buildings relates to the discrepancy between how buildings are designed/expected to perform and how they actually perform. The discrepancy is known as the building performance gap. The performance gap is the discrepancy between a building’s expected and actual occupant satisfaction or energy use. Generally, the building performance gap is attributed to the limited information on buildings’ operational performance during building design. This information can be obtained by collecting and analysing operational data during building operation using data-driven technologies such as the Internet of Things and machine learning. Current research focuses on the application of these technologies primarily from a technical point of view. However, building actors lack feasible methods for collecting and analysing building operational data to reduce the building performance gap. The overall research objective of the PhD thesis was to obtain insights into the actual operation of buildings for the benefit of occupant satisfaction, energy efficiency and building design. For this purpose, four data-driven methods were developed addressing the following aspects: Occupant thermal comfort, indoor air quality, energy efficiency and heating ventilation and air-conditioning (HVAC) sizing. The overall success criterion was that the methods provided insights that would be used during building design or operation to limit the building performance gap. The data-driven methods were applied to different use-cases in a Danish office building to investigate if they provided insights into a building’s design and operational performance. The thesis demonstrated that long-term measurements of airflow, electricity use, indoor temperature and CO2-concentration provided insights into the operational performance of air-handling units and showed that these units were sized appropriately according to the design requirements in the studied building. Passive-infrared sensors mounted below the building occupants’ work-desks showed that the expected occupancy level of 100% overestimated the observed occupancy level of 70% in the studied building. The application of continuous thermal feedback revealed that the studied building’s occupants preferred an indoor temperature level of 23.5 °C. Furthermore, prediction models of the studied building suggested that if the obtained insights were implemented in the studied building, the building occupants would obtain an increased thermal comfort of up to 10%, and the building would save up to 40% in energy use. In conclusion, the results of the thesis showed that the proposed data-driven methods provided insights into buildings’ actual operational performance to benefit occupant satisfaction, energy efficiency and building design. The insights above should be used in the particular building to improve the current building operation. Moreover, they should be used indicatively in building design, thereby opening up for an evidence-based design approach.ac |
