Abstract
We developed a neural network model describing the process in which disturbance conditions, such as outdoor air temperature, and activity conditions, such as indoor air temperature and humidity, affected the chilled water thermal quantity for HVAC systems in a commercial building in the city of Kitakyushu, Japan within three years after its completion. The influence of each of the disturbance conditions and the activity conditions, which changed every year, on the chilled water thermal quantity was diagnosed quantitatively by a numerical simulation using these models, and how to macroscopically evaluate the carried-out energy-saving activity was shown.
