Thermal comfort and energy efficiency evaluation of air-conditioning systems using machine learning

• Main Author: Muhammad Fahmi Bin Ali
• Other Authors: Soh Yeng Chai
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/67421
• Institution: Nanyang Technological University

Air-conditioning systems are widely used around the world and it gives a comfortable atmosphere and thermal comfort (TC) towards its occupants. However, out of the total energy consumption used, 50% comes from only the air-conditioning systems in a commercial building, It is found that TC plays a part in this and finding the relationship between that of thermal comfort and air conditioning system (A/C) as the primary objective. TC relies heavily on environmental and physiological features such as MRT, RH, TA, thermal insulation of clothing, AV, and human activities. However, thermal comfort is a subjective matter because every individual perceive thermal comfort differently. Therefore, it is crucial to evaluate and investigate the ranges of acceptable temperature that will suit the majority of the occupants. However, according to the ASHRAE, it is considered healthy and acceptable when an indoor environment reaches or when it is at least 80% satisfied based on the occupants satisfaction with the thermal comfort levels in the building. This project uses machine learning techniques as the methodology in finding the objective. Techniques that will be using in this project will be those such as SVM and ELM. These algorithms have proven its use towards finding nonlinear and complex systems through successful applications in medical and industrial fields. Although ELM is can be considered a new type of machine learning technique, it has been gaining popularity for its outstanding performance in terms of learning speed and adaptability.