Development of a computational tool to quantify architectural-design effects on thermal comfort in naturally ventilated rural houses

In the present study, the effect of the opening size and building direction on night hours thermal comfort in a naturally ventilated rural house is investigated. Initially, the airflow in and around the building is simulated using a validated computational fluid dynamics (CFD) model. Local climate night-time data (wind velocity and direction, temperature and relative humidity) are recorded in a weather station and the prevailing conditions are imposed in the CFD model as inlet boundary conditions. The produced airflow patterns are then used to evaluate indoor thermal comfort. For this reason, special thermal comfort indices, i.e. the well-known predicted mean vote (PMV) index and its modifications especially for natural ventilation, are calculated with respect to various residential activities. Mean values of these indices (output variables) within the occupied zone are calculated for different combinations of opening sizes and building directions (input variables), to generate a database of input–output pairs. Finally, the database is used to train and validate Radial Basis Function Artificial Neural Network (RBF ANN) input–output “meta-models”. It is demonstrated that the proposed methodology leads to reliable thermal comfort predictions, while the optimum design variables are easily recognized.