With progress in information technology, the controllability of the building facility system continues to improve. Especially, unlike the heat source air-conditioning equipment, the lighting equipment has little time lag, so fine control over time and space is possible. Recently, it has become technically possible to allocate an IP address to each lighting fixture and independently control all the lighting fixtures. To evaluate the value of such advanced control, it is necessary to check whether the control works effectively for human activity in the building. However, if it is impossible to judge whether good or bad is necessary without measuring the real building, it is inconvenient to design the control. Therefore, using a simulation, we aimed to predict the energy-saving effect of the lighting control due to the absence of workers in the office.
We divided the office space into meshes and constructed an agent model of the office worker moving between meshes using the A* algorithm. By using several office worker models, we calculated the time the worker is in each office's mesh. We controlled the lighting fixtures in this virtual office room based on human sensors. We quantitatively evaluated the influence of such conditions as the occupancy rate, waiting time, fade-out time, and minimum output rate on the energy consumption rate of lighting fixtures. The energy consumption rate was roughly consistent with the survey results in past research. An approximation formula was developed to predict energy consumption rate, and the procedure for predicting the energy-saving effect of human sensor control is described.