Description
Reducing environment impact by tackling challenges in energy management has become a central topic for highly urbanized cities such as Hong Kong. One major source of carbon emissions comes from the generation of electricity, of which 90% of this generated electricity is consumed by buildings.The Electrical and Mechanical Services Department (EMSD) of the HKSAR Government is managing over 8000 government facilities. In view of various electrical and mechanical equipment, Heating, Ventilation and Air Conditioning (HVAC) system is consuming most of the electricity inside a building. To optimize the energy efficiency of a HVAC system, EMSD has made use of artificial intelligence (AI) technologies to dynamically configure the operations of different chillers in a HVAC system based on the cooling demand. The AI model was tested in a government building and was able to achieve an accuracy of 99% in forecasting the cooling demand.
To extend the AI model to other buildings in the city, we have developed an innovative approach using Semantic AI to scale gains from single building to a city level. Semantic AI is a combination of knowledge graphs, natural language processing and artificial intelligence. It is a unified semantic model that encapsulated the context of a system and the representation of the semantic model is in a machine-readable format, enabling programmatic exploration of different operational, structural and functional facets of a building.
In this paper we will present the challenges faced and the experience gained during the implementation of Semantic AI. We will also discuss how the development of a semantic AI model will accelerate the optimization of energy efficiency of HVAC systems in Hong Kong, and thus achieving carbon neutrality before 2050 as targeted by the HKSAR Government.
Number of attendees (for events)
200Period | 23 Nov 2022 |
---|---|
Held at | Institute of Hong Kong, Hong Kong |
Degree of Recognition | Regional |
User-Defined Keywords
- Carbon Neutrality, Semantic Model, Ontology, Machine Learning, Smart Building