Building occupancy significantly impacts energy use, timing for demand impacts, and is a significant source of uncertainty in building energy models. There are relatively few sources that define building occupancy schedules and number of occupants per building or space type. More importantly, these sources define traditional schedules that are likely not to reflect the true occupancy of a given building. We construct traffic-based occupancy schedules which are more responsive to changes in mobility patterns, and which can realistically estimate occupant arrivals, departures, and counts in individual buildings.
In Proceedings of Building Simulation 2021: 17th Conference of IBPSA
An overview of the methodology presented in this paper, guided by the data workflow. The left-hand side of the figure, (a)-(d), illustrates the traffic data workflow: the data is collected from sensors, and connected through a topological model. Based on vehicle turning movements, we determine the number if vehicles entering the link between two intersections, and the vehicles leaving the link. The difference in vehicles is assigned to the two intersections, relative to their available building space. The vehicle counts are multiplied by 1.67, the average number of occupants per vehicle, to reflect the number of prospective building occupants. The right-hand side of the figure, (f )-(h), illustrates the building workflow: each building is assigned to all nearby intersections, based on their Voronoi cells. If the building footprint intersects with multiple cells, the building area is assigned proportionally to the different intersections based on the amount of overlap between footprint and Voronoi cell. We eliminate buildings which are outside reasonable walking distance. Finally, the two workflows join at the bottom of the figure (e): the prospective building occupants for each intersection are distributed to its assigned buildings (partial or full) based on the available building area.
