An Air Balancing Method Using Artificial Neural Networks for the Ventilation System

Abstract

Physical model of ventilation system requiring exact knowledge of all component parameters is usual unavailable in practice. Based on this issue, the objective of this study is to develop an air balancing method to predict the damper position provided the desired airflow rate without requiring a physical model. In the study reported here, the proposed air balancing method consists of a multilayer perceptron model and a damper control method. The multilayer perceptron model is constructed and trained to simulate the non-linear relationship between pressure drop and airflow rate at the damper, and the damper position control method is used to relate pressure drop to operating position of the damper. Experimental tests are carried out to validate the performance of the proposed method. The results show that the proposed method is powerful to balance the ventilation system.