Minimum number of simulation runs for reliable building energy and peak load prediction at different building scales: a study on stochastic shade adjustment

This paper investigates the impact of uncertain shade behavior on three facades (east, south and west) on energy performance. A previously developed stochastic shade behavior model and an occupancy model were used to conduct building co-simulation for determining energy-related performance uncertainty (annual energy demand and peak heating/cooling loads) at different building scales. Statistical test and Monte Carlo random sampling techniques were used to determine the required minimum simulation runs for obtaining reliable annual energy performance and peak loads. The results show that there is no significant difference in shading performance among seasons and orientations. In China, most newly-built office buildings are high-rise ones and thus, the influence of manual shades on the energy uncertainty of these office buildings can be neglected and a single simulation run is enough to obtain reliable annual heating, cooling and total energy demands. While for peak heating/cooling loads, the conclusion is different from annual energy demand and at least five simulation runs are required to obtain peak heating/cooling loads at building scales. For a smaller building, 120 simulation runs are suggested for peak load determination.