One of the key purposes of scientific researches and inventions is to make our life better. For example, air conditioning system is supposed to make us feel cool and comfortable in summer. But feelings are subjective and one man’s comfortable temperature could be too low for others. In Hong Kong, most people working for prolonged hours in offices with central air conditioning find the temperature on the low side. This incurs tremendous cost and energy wastage while sacrificing the comfort of the users. In light of this, Dr Dan Wang, Associate Professor, Department of Computing, led a research team in devising the Scalable Personalized Thermal-Comfort Platform (SPET) that allows users to vote on the temperature they prefer. In a test trial, the platform helped save 18% of energy and improve user thermal comfort by 33.8%.
Physiological thermal comfort model
“Smart” has been one of the most mentioned tech buzzwords in our daily life. Our phones are smart. Our cars are smart. Our homes are smart and of course, a city can only be smart if most of its skyscrapers are smart. One criterion to judge whether a building is smart is its energy efficiency. According to a report by EMSD, energy consumed by the commercial sector accounted for 65% of Hong Kong’s total energy consumption in 2014. Half of that energy went to air conditioning. In other words, to conserve energy, starting with air conditioning makes a lot of sense.
Dr Wang believed most building managers would rather set the temperature lower than needed because users may complain if it’s not cold enough. “For decades, researchers have been suggesting an office temperature between 21 to 23°C to optimize productivity. But user comfort is subjective and we believe a platform that lets users express how they feel would be useful,” said Dr Wang. In 1970s, researchers started publishing papers on the physiological thermal comfort model, outlining the factors that influence people’s feeling about a certain temperature. “Generally speaking, a person would feel hot when the body generates more heat than the heat loss. Personal factors such as metabolic rate, sex, weight, body fat level, age and clothing all affect thermal comfort. So are other external factors. We tried to factor them in as far as possible.”
A voting system for indoor temperature
The Scalable Personalized Thermal-Comfort Platform (or SPET in short) collects personal information of the users, such as weight, height, sex and BMI, via a smartphone app. Then users may vote whether the existing temperature is perfect, too cold or too warm. “The platform also collects external data such as the time, outdoor and indoor temperature, fan speed and ultraviolet index. Such figures alongside the personal data of all users would be incorporated in a formula to calculate the best indoor temperature that most users are comfortable with. The air conditioning system would then be set at that temperature,” explained Dr Wang.
However, not all users in an office would vote. In such cases, the platform has a cross-learning algorithm to make the result more representative. “The platform can learn from past records to deduce the user’s preference on a similar day at the similar time. Besides, the BMI of the user not voting is used as a basis and cross reference to preferences among other voting users of the similar BMI is made for calculation,” he added.
In a three-week trial in a 484-square-foot office with 13 workers, users expressed a level of thermal comfort increased by 33.8% while 18% less energy was consumed on air conditioning. Average indoor temperature was raised 1.75°C. The platform can be easily retrofitted on existing central air conditioning system without large-scale modification.
