|Abstract:||Designers develop high-performance office building designs with efficient envelopes, and Heating, Ventilation and Air Conditioning (HVAC) systems utilizing energy modeling tools with forecasted occupancy, plug load, and operational profiles as inputs. Their expected performance is often not met due to discrepancies between modeled and actual energy behaviors resulting in energy performance gaps. Energy performance gaps stemming out of design, construction and operation phases, pose serious challenges to the credibility of the design and engineering sectors in achieving predicted energy goals of the project.
Data-driven models, based on the actual building energy consumption data, offer an excellent means to evaluate significant root-causes of energy performance gaps; occupancy, envelope and HVAC operations. On the one hand, this supports the feedback loop to the design process from real-time building energy performance to predict energy performance accurately, choose optimal design solutions and, mitigate construction quality management issues in future designs. On the other hand, this provides perfect feedback to optimize energy performance in existing buildings.
This doctoral research presents data-driven modeling methodologies leveraging energy consumption data, advanced statistical methods, and expert domain knowledge, to assess the occupancy profiles, envelope thermal response, and HVAC systems’ performance, in office buildings. The study proposes data-driven techniques to generate evidence-based knowledge to designers about the real-time energy performance of their design decisions and required inputs for energy simulation models, that support the development of high-performance office designs in the future with minimal energy performance gaps.