This post is part of a series on occupant modeling in CEA. In previous posts, we have explored the CEA deterministic and stochastic occupancy models. Since both of these models are based on standard occupant schedules for individual occupancy types, they share the underlying assumption that any building of a given occupancy type will be the same as any other building of the same occupancy type.
However, we know this to be an oversimplification. A recent study in our team compared the standard occupancy patterns to data available online on so-called “popular times” to attend specific retail and restaurant buildings in different cities. These results show that on average these schedules are off by 10–20% from reality! It becomes evident that actual building occupancy patterns can greatly differ from the standards.
In a different study, we have also quantified the variation caused by the choice of occupant model by looking at more than 100 buildings in an urban area in Zürich and observing how much the final energy demands varied from the standard schedule when using the stochastic schedule described in the previous blog entry and actual data of occupant presence in these buildings. The results showed that the deviation on a yearly basis is relatively low for all energy demands considered, which might explain why historically it has made sense to use simple approximations of occupant presence. However, with increasing time resolution the results become less reliable, and the peak demand (which is important for supply system sizing) showed an average variation of 15% for cooling and electricity for lighting and appliances.
With the increasing availability of data about buildings and their occupants, we can now make increasingly robust predictions of occupants’ presence in buildings and the associated demands they create. It therefore makes sense to base these assumptions on actual data for individual buildings rather than on general schedules for the given occupancy types.
In upcoming versions of CEA, users will be able to define occupant schedules for individual buildings and for individual occupants when such data is available, or revert back to standard schedules when in need of general estimates due to lack of data.
We’ll keep you posted!