The modeling activity presented in this work aims at the assessment of a simplified model, named BS model, which was specifically developed for integration of DSF in Building Simulation. The BS model is based on a pressure loop and on an integral approach to the heat transfer along the vertical channel. It considers buoyancy as a function of the average temperature in the channel. The wind action is taken into account by means of wind pressure coefficients (Cp) on the facade openings. The focus of this study is the experimental validation of the modeling "core": the natural ventilation through the DSF. The validation is based on the dataset of the experimental campaign conducted on a DSF test facility, the "Cube", in Denmark, under IEA ECBCS ANNEX 43/SHC Task 34. Hourly simulations were performed with the BS model for the 15 days of the experimental campaign. A CFD modeling activity was also carried out on a selection of four cases, extracted from the experimental benchmark and representative of different temperature and pressure boundary conditions. The results show that the BS model presents a good level of agreement with the experimental data in predicting the mass flow rate and the heat removed by ventilation. Although the two experimental methods used to determine the airflow rate in the DSF cavity produce in many cases divergent results, it was possible to distinguish valid experimental results for comparison with the BS model. This was possible thanks to a thorough analysis of the experimental procedure together with the insight provided by the model into the determination of the driving wind and thermal differential pressures. In particular, by selecting only the measurements associated to sufficiently low wind fluctuations in the hourly averaged data, a good degree of correlation was found between the predicted total driving pressure and the flow measurements. Concerning the four cases investigated also by means of CFD, the agreement between the BS and CFD models is remarkable in terms of outlet temperatures and in the prediction of flow reversal. (C) 2017 Elsevier B.V. All rights reserved.