An energy-saving control strategy for multi-zone demand controlled ventilation system with data-driven model and air balancing control

Abstract

A data-driven energy-saving control strategy applied to balance the multi-zone demand controlled ventilation system is presented. The proposed strategy consists of two steps: system model construction and air balancing control. Based on observed datasets, a multi-layer perceptron structure is employed to model the multi-zone ventilation system. The model is used to predict the pressure differences of each damper based on the static pressure of the main duct and the desired airflow rates of each damper. Air balancing control approach is implemented based on the empirical formula of the damper. This approach is use to predict the operating positions of each damper based on the predicted pressure differences of the developed model. An experimental apparatus consisting of original components of ventilation system is set up to collect the training and testing data, and simultaneously used to validate the performance of the proposed control strategy. Experimental results demonstrate that the issue of over-ventilation and under-ventilation of demand controlled ventilation system is eliminated, and energy savings of fan power can be obtained with the proposed control strategy.