Neurodynamics-driven Optimization and Control of Intelligent Heating, Ventilation and Air Conditioning Systems

Description

Heating, ventilation, and air conditioning (HVAC) systems are vital facilities for regulating the indoor environment in residential, industrial, and commercial buildings to meet specified human comfort and air quality requirements. As reported by the Hong Kong’s Electrical and Mechanical Services Department, HVAC systems consumed 26% of energy in commercial-residential sectors of Hong Kong. Globally, HVAC systems account for about 20% and 10% energy consumption in USA and the whole world, respectively. In the global urbanization process, it is anticipated that HVAC systems will take up an increasing portion of energy consumption. In contrast to conventional HVAC systems with a few pneumatic, thermal, and electrical devices, modern HVAC systems in large buildings consist of hundreds to thousands of multi-modal devices. With the advances in the Internet of Things (IoT) technologies, many such devices are equipped with computation and communication capabilities which facilitate the optimization and control of HVAC systems for load dispatching, system monitoring, and control. In view of the available IoT technologies and global demands for improving living quality and reducing energy consumption and carbon emission, it is both technically feasible and economically beneficial to develop intelligent HVAC systems in building automation of smart cities. As a backbone of artificial intelligence, neural computation attracts significant attention in recent years due to its successful applications for modeling, optimization, and control of many intelligent systems. In particular, our recent research results on collaborative neurodynamic approaches to global optimization and intelligent control demonstrated their great potentials for applications in HVAC systems. In the proposed research, a systematic investigation will be carried out on the optimization and control of HVAC systems based on neural computation. The research will consist of four coherent parts. The first part will focus on developing neurodynamic approaches to optimal senor-actuator deployment in networked HVAC systems. The second part will focus on developing collaborative neurodynamic approaches to distributed load dispatching and coordination of multi-modal HVAC devices under various operational constraints. The third part will focus on developing intelligent methods based on neural networks for modeling and control of IoT devices and HVAC systems subject to various disturbances. The fourth part will focus on validating the efficacy of the proposed neurodynamics-driven methodologies in regulating ambient thermal states and improving energy efficiency in a cyber-physical system in various settings. It is expected that the successful completion of the proposed project will significantly advance the research and development of intelligent HVAC systems.