Procedures for scaling and replication by simulation

The most effective way of establishing confidence in the ability of a simulation tool to model a particular component or system is to compare the predictions with measured data. These data can derive from experiments in test cells or monitored test buildings. In both cases, however, the performance cannot be directly translated to generalised conclusions because the conditions of the test are not representative of conditions within typical real buildings. A procedure has been developed using simulation which permits the scaling of performance of a building component on a test cell to the performance of the same component on a real building. It involves the selection of a reference building which is representative of a particular class of building and for which representative conditions of occupancy, infiltration, ventilation, control and climate are specified. It is believed that this procedure is more flexible and able to deal with complex interactions when compared with the alternative procedure which involves the extraction of performance characteristics from the experimental data and their application using a simplified design tool.