Impacts of uncertainty in energy modelling widely used in aggressive energy efficiency regulations

Building Assessment Tools (BATs) are widely used to estimate the performance of building and to assist designers in making decisions. As building codes and rating systems move from prescriptive to performance-based metrics, BATs are increasingly used to show compliance. BATs use computational methods and the results are mostly in a single annualised metric. However, the scientific community has shown that aleatory factors such as occupant behaviour and weather make the potential energy use of a building far from being a single deterministic value. Also, it is known that there is a significant deviation between predicted (at design stage) and actual energy use in buildings. These variations reduce the credibility of the predictions, questioning the acceptance of BATs results without considering underlying errors. This problem is amplified in under-policed construction sectors. India is one such example, where the adoption of Energy Conservation Building Code is becoming mandatory and one of its compliance method is based on BATs. Our work, therefore looks at the uncertainty in a typical multi-storey commercial office building's performance in Delhi and shows implications of variable inputs in the results. The paper first reviews the use of BATs and existing studies on simulation uncertainty. Then uncertainty is evaluated in energy simulation of a sample building, including effects of inconsistent construction practices. EnergyPlus is then fed values sampled (by Monte-Carlo method) from probability distribution functions of inputs (building fabric and operational parameters). Further sensitivity and uncertainty analysis of the results is performed. From the 3500+ simulations, the most sensitive inputs found were Wall U-values, cooling setpoints and infiltration. The variation in cooling demand and discomfort hours is more than three times between best and worst cases.