Modeling city-scale building energy dynamics through inter-connected distributed adjacency blocks

Buildings consume the largest amount of energy in cities and simulating urban energy dynamics provides the most cost-effective references for urban building planning and energy policy-making. However, cities have a tremendous number of buildings and complicated physical/environmental conditions, current simulation models require formidable computation resources and time. This paper proposes a rapid simulation approach that decomposes city model into spatially correlated building blocks for distributed simulation. The proposed distributed adjacency blocks (DABs) algorithm utilizes 2D footprint to construct 3D building groups and solar azimuth angles, altitude angles, and shading plane to simplify simulation targets. With the proposed method, the energy dynamics of the whole city can be simulated in parallel with multiple threads through abstracting inter-building boundary conditions. To validate the proposed method, this study conducted two validation experiments with different building numbers, window-to-wall ratio, and climate conditions. The simulation results suggested that the proposed algorithm can dramatically improve the simulation efficiency and generate less than 5% of percentage difference compared with the conventional whole city simulation approaches.