Prediction of air flow in a single-sized naturally ventilated test room using artificial neural networks.

The objective of this research is to investigate thermal comfort and air flow distribution insidea test room which is naturally ventilated. The test room is ventilated through adjustablelouvers. The air pressures and velocities across the openings together with indoor airtemperature and mean velocity at four locations and six different levels are measured. Thecollected data are used to predict thermal comfort parameters across the test room. Tests werecarried out over the winter and summer time. The experimental results demonstrate that forboth winter and summer the air was entering the test room at bottom and leaving at the toplouver. The amount of air flow over the winter was increased as the heat input varied from 2kW to 4 kW. The predicted thermal comfort indicated that the Percentage of PeopleDissatisfied (PPD) values for summer are significantly improved with a higher temperaturedifference between inside and outside and higher wind velocity.