Air balancing technology in HVAC applications

• Main Author: Cui, Can
• Other Authors: Cai Wenjian
• Format: Theses and Dissertations
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation
• Online Access: https://hdl.handle.net/10356/103469; http://hdl.handle.net/10220/48071
• Institution: Nanyang Technological University

This thesis presents a detailed investigation into the air balancing process in HVAC systems. Improvements have been made in both the flow measurement and air balancing methods to optimize the balancing performance in terms of efficiency, accuracy and cost. The main contributions of the thesis are summarized as follows: The accuracy of flow measurement in the air flow control stations (AFCS) is improved to achieve better control performance of the air balancing methods. A revised airflow rate calculation formula for the averaging Pitot tube is proposed taking into consideration of downstream disturbances. The revised formula significantly reduces the systematic error caused by the flow control damper and improves the accuracy of flow measurement to a large extent. Besides, the AFCS is equipped with wireless technology to achieve wireless communication. Two online air balancing methods are proposed, which are distributed cooperative control-based air balancing (DCC-AB) method and gradient-based online adaptive balancing (GOAB) method. Both methods can be performed during system operation without interrupting the normal service of ventilation, which avoids inconvenience to the occupants and financial losses caused by maintenance shutdown. The balancing-in-service characteristic provides possibility to achieve accurate air supply under dynamic loads, which indicates a large energy saving potential. The DCC-AB method is based on the distributed cooperative control strategy. It guarantees an asymptotic convergence towards the balanced state. the balancing process starts from the given imbalanced state and converges asymptotically to the balanced state. The drastic change of damper positions such as short cycling on and off is circumvented, which avoids overshooting and oscillations of the flow rate value. An incremental improvement in the imbalance of flow is guaranteed at each adjustment during the whole process. The DCC-AB method circumvents the centralized supervisory control and requires only a sparse communication architecture to cooperatively achieve the objective. The method also incorporates an additional adjustable term that reflects the total power consumption to optimize the energy efficiency. Besides, this method is a model-free method that requires little prior knowledge on the system topology and duct parameters. The GOAB method is based on the gradient descent algorithm. In this method, an objective function is defined to quantify the discrepancy between the normalized flow rate and the set point, and the increment of damper angle is calculated using stochastic gradient descent. An online adaptive mechanism for the Jacobian estimation is applied to capture the change of Jacobian matrix during the balancing process. In this method, the critical damper is explicitly identified based on the null space of the gradient vector. At least one damper is guaranteed fully open, which minimizes the overall flow resistance of dampers and total energy consumption. The GOAB method incorporates a coefficient to control the speed of convergence rate and stability of the algorithm. A rapid and stable convergence is guaranteed within several steps.