Modeling night-time ventilation in Stanford’s Y2E2 building

An important fraction of the energy consumption in buildings is related to controlling the indoor temperature. In moderate climates passive cooling by nighttime ventilation can result in significant energy savings, but modeling natural ventilation systems for design and performance predictions remains a challenging task. We investigate the predictive capability of a box model and a CFD simulation for modeling night-flush ventilation in the Y2E2 building on Stanford’s campus. Both models consider a single atrium of the building and the predictions are compared to available temperature measurements. The box model solves for the average air and building thermal mass temperatures, and represents heat sources and sinks as integral values. The uncertainty in the input parameters for the box model is propagated to the output using a non-intrusive polynomial chaos method. The mean result predicts a too fast cooling rate, but the maximum discrepancy with the measured data is limited to 0.6K, and the measurements are within the predicted 95%CI. The uncertainty in the results is mainly related to uncertainty in the infiltration flow rate, the initial thermal mass temperature, the internal heat source, and the heat transfer coefficient. The CFD simulation represents a much higher level of detail in the building model, but it also predicts a too high cooling rate. The maximum discrepancy between the average air temperature from the CFD model and the measurement was found to be 0.9K. The discharge coefficients computed from the CFD simulation depend on the floor and the time during the nightflush and vary from 1.1 to 0.15, which is outside of the range originally assumed in the box model UQ study. The average heat transfer coefficient computed from the simulations is around 2.8W/mK, which is within the range used by the box model.