Artificial neural networks for predicting the pressure coefficients in a naturally ventilated test room.

Describes an investigation of the possibility of using artificial neural networks to predict air pressure coefficients across openings in a lightweight single-sided naturally ventilated test room. The network was trained using experimental values. A monitoring experiments was carried out on the outside local temperature, wind velocity and direction. Estimates were made of the pressure coefficients at the top and bottom of the openings from the recorded data of air pressures and velocities across the openings with indoor air temperatures. The collected data and the air heater load and a factor indicating whether the opening is situated to the windward or leeward direction were used as input to the neural network and the estimated pressure coefficients were used as the output. With one hidden slab a general regression neural network was used. Satisfactory accuracy and correlation coefficients were obtained for the two coefficients respectively, and satisfactory results were obtained when unknown data were input.