CALIBRATION AND MODELING FOR A DASHBOARD THAT PROVIDES REAL- TIME FEEDBACK ON ENERGY SAVING STRATEGIES

A new approach toward using building dashboards to educate building users is proposed, and a modeling framework tested for a sample building. The goal was to create a methodology that allows the user to select from a variety of potentially energy saving actions or green building operation strategies and see how the hypothetical change would affect the building’s energy consumption both at the current moment and over the year. Neural networks, trained with extensive stochastic runs of a calibrated Energy Plus model, provide a quick estimate of the instantaneous savings or cost of implementing a given scenario, with a level of calculation intensity that can be completed repeatedly on a tablet or a server serving multiple tablets. The error was tested with a year of simulation data, and each scenario demonstrates less than 6% bias error in the energy savings.