Transient simulation calibration of an old building using an experimental design: evaluating uncertainty results

Calibration of Building Energy Simulation is an under-determined problem. There can be several correct solutions. The consequences of using the Building Energy Simulation for retrofit design can be significant. In this paper, we will propose a method that identifies different solutions in the entire field that was studied. It enables one to evaluate the uncertainty in energy saving estimations. The method uses an experimental design in order to reduce calculation time. It uses the coefficient of variation of root mean square error. We applied the method on an old building. The same retrofit design can yield a 40 to 60% savings in energy according to the chosen calibrated Building Energy Simulation.