Modeling, air balancing and optimal pressure set-point selection for the ventilation system with minimized energy consumption

• Main Authors: Jing, Gang, Cai, Wenjian, Zhang, Xin, Cui, Can, Yin, Xiaohong, Xian, Huacai
• Other Authors: School of Electrical and Electronic Engineering
• Format: Article
• Language: English
• Published: 2021
• Subjects: Engineering::Electrical and electronic engineering; Ventilation; Air Balancing
• Online Access: https://hdl.handle.net/10356/150547
• Institution: Nanyang Technological University

Traditional static pressure reset control strategies commonly use a feedback indicator to reset the static pressure; this results in under-ventilation in certain zones and over-ventilation in others. Based on this issue, the objective of this study was to develop a model-based, improved, static pressure reset control strategy, providing a well-balanced system to eliminate under-ventilation and over-ventilation, while consuming minimal energy. In the study reported here, a comprehensive mathematical model was established to simulate the non-linear behavior of the ventilation system, and a supervised machine learning algorithm for a support vector machine was used to obtain values for unknown parameters in the model. The resulting model was then used as the basis for development of a damper position control method and to determine the damper position, given a desired airflow rate. An optimal, static pressure set-point selection method was also proposed using the developed model to calculate the minimum static pressure set-point in a closed-form. As a result, the revised system consumed less energy owing to the better-balanced system and optimized pressure set-point selection. Moreover, through the application of the damper position control method, the ventilation system was well-balanced and eliminated both under-ventilation and over-ventilation. Experimental tests were carried out to validate the performance of the proposed method in comparison with the conventional static pressure reset strategy, data from which were collected to train the proposed model.