Performance monitoring in buildings involves tracking variations in temperature sensitivity and baseload over time to evaluate cooling consumption patterns and identify opportunities for retrofitting and optimization.
Buildings that exhibit significant fluctuations in their performance may indicate inefficiencies or inconsistencies in their operation. While occasional variations can be expected due to changes in external factors like weather or occupancy, excessive jumping around could suggest potential issues that need to be addressed.
Buildings that consistently exhibit higher temperature sensitivities could benefit from window upgrades, better insulation, and installation of awnings. On the other hand, reduction in cooling of these buildings is likely to make their occupants uncomfortable during heat waves.
"Drivers" in the context of cooling consumption in buildings refer to the key factors that significantly influence the amount of cooling consumed. These drivers play a pivotal role in understanding and predicting cooling usage patterns within buildings.
We consider various weather-related factors including, Outside Air Temperature, Relative Humidity (RH), Dew Point of Temperature, Wind, and Solar Radiation. We also use Weekend Indicator as a proxy for building occupancy, indicating whether it is a weekday or weekend. This helps gauge the influence of building occupants' behavior on cooling use.
Mean daily outside air temperature is identified as the key driver for cooling consumption. Alone, it explains over 70% of the variance in cooling load for buildings that represent 82% of portfolio-level load. The weekend indicator is the second most important feature, followed by relative humidity.
Cooling demands in commercial buildings vary significantly, both among individual buildings and within a specific type of buildings. These differences are mainly driven by the building's usage and exposure to external environment.
The bubbles represent different buildings and their colors indicate building types, while their sizes indicate the floorspace of buildings. This chart associates the temperature sensitivity and base load intensity of buildings to building types.
Buildings with labs have varying temperature sensitivity and base load intensity due to differences in air changes and use of process loads. Healthcare buildings maintain a stable indoor environment with low temperature sensitivity and high base load intensity, while Offices/Classrooms show significant variations due to differing occupancy and usage.
We have two types of models: Linear and Log-Linear. Linear models directly relate cooling demand to temperature, while Log-Linear models first take the logarithm of cooling demand. The advantage of Log-Linear models is that they let us express changes as percentages.
In the chart below, each bar represents a different building. The height of the bar shows how well the model performs (measured by R²), and the color indicates the type of building. The x-axis represents the building's share of the total load.
These models generate higher quality predictions for most Health care and Labs buildings i.e., these buildings are well-explained, while shorter bars are observed for majority of offices and classrooms buildings.